There appear to be two fundamentally different approaches to data analytics: one which seeks to identify the latent content of communications through statistical methods (using information theory, Baysian analysis, etc); and the other which seeks to create shared contexts for the exploration of data. The context within which both these techniques aim to intervene is the highly-charged political environment of businesses and institutions. The reasons why there is so much attention being paid to the topic is because of the impact of information and the technologies which generate it on the working practices, governance and viability of businesses and institutions (including of course, educational institutions!)
Information in an institutional context can be seen in a variety of ways. In my thinking about games and metagames in institutional politics (from Nigel Howard's work), I have been trying to decide whether information provides the context for making decisions (as traditional game theory would suggest), or whether in fact information is in some way a 'player' generating potentially unpredictable 'moves' to which all the other players have to react. Certainly in thinking how politicians can get 'caught out' by a set of 'bad statistics' suggests that information indeed can appear like a 'player', tripping up whoever has been in the game of making assertions that the information's "move" denies. In fact, when we play Monopoly, much of the 'information' which changes the game is contained in 'chance' card or the roll of the dice.
If the objective of a political game concerns the maintaining of coalitions, then the ways in which information is conceived within that overall game is important. One thing that can be done is to try to predict the "roll of the dice", so that plans can be made for the maintenance of the coalition on the basis of some analysis of what information might throw up. It may be that data analytics may provide indications as the what the information might contain next (certainly, it isn't a random entity - it's behaviour is more like a Markov series). However, the problem with an anlytical approach in a game context is that the analysis itself is a move in the game: players have to be convinced of the rationale for the analysis, significant intervention is required to gather the right data (and decide on what the right data is), and resources need to be put in place to analyse that data in the right way. The data analytical game is a meta-game, but one which may serve to divide allegiances rather than unite people. The payoffs of the data analytical game may not, in the final analysis, be sufficient for it to be truly realised.
But there is another way and one which I think might carry more weight. This is the 'activity' route. Or rather, the 'activity+technology' route. Information in this situation is both context and player. Technology can be used to amplify the information environment as a shared context for an activity, or a kind of artificial game, the purpose of which is to reveal deeper understanding into the meanings of and explanations for the information between each individual who engages with the activity. Unlike the data analytical approach, the activity approach assumes that the meaning of data is in the people who examine it, and not the data itself. It puts emphasis on the way things are explained. But more importantly, it focuses on the key goal of the games as being the maintaining of coalitions, and rather that trying to predict the next "move" of information as a vehicle for maintaining coalitions, it creates an artificial context for the building of a coalition and through that context, allows individuals to explore and express the meaning they see in the information they encounter.
There are many things that fascinate me about this latter approach. Not least, because it highlights the importance of togetherness and convivial experience in the identification of meaning and the coordination of action. It also strikes me as interesting that such an artifical activity context looks remarkably like Beer's Syntegrity (see http://www.syntegration.com/). Integrating such approaches with a deep understanding of information, and the political games that are played in institutions produces a fascinating practical approach which addresses some of the problems that institutions face in a rapidly moving information environment.
Information in an institutional context can be seen in a variety of ways. In my thinking about games and metagames in institutional politics (from Nigel Howard's work), I have been trying to decide whether information provides the context for making decisions (as traditional game theory would suggest), or whether in fact information is in some way a 'player' generating potentially unpredictable 'moves' to which all the other players have to react. Certainly in thinking how politicians can get 'caught out' by a set of 'bad statistics' suggests that information indeed can appear like a 'player', tripping up whoever has been in the game of making assertions that the information's "move" denies. In fact, when we play Monopoly, much of the 'information' which changes the game is contained in 'chance' card or the roll of the dice.
If the objective of a political game concerns the maintaining of coalitions, then the ways in which information is conceived within that overall game is important. One thing that can be done is to try to predict the "roll of the dice", so that plans can be made for the maintenance of the coalition on the basis of some analysis of what information might throw up. It may be that data analytics may provide indications as the what the information might contain next (certainly, it isn't a random entity - it's behaviour is more like a Markov series). However, the problem with an anlytical approach in a game context is that the analysis itself is a move in the game: players have to be convinced of the rationale for the analysis, significant intervention is required to gather the right data (and decide on what the right data is), and resources need to be put in place to analyse that data in the right way. The data analytical game is a meta-game, but one which may serve to divide allegiances rather than unite people. The payoffs of the data analytical game may not, in the final analysis, be sufficient for it to be truly realised.
But there is another way and one which I think might carry more weight. This is the 'activity' route. Or rather, the 'activity+technology' route. Information in this situation is both context and player. Technology can be used to amplify the information environment as a shared context for an activity, or a kind of artificial game, the purpose of which is to reveal deeper understanding into the meanings of and explanations for the information between each individual who engages with the activity. Unlike the data analytical approach, the activity approach assumes that the meaning of data is in the people who examine it, and not the data itself. It puts emphasis on the way things are explained. But more importantly, it focuses on the key goal of the games as being the maintaining of coalitions, and rather that trying to predict the next "move" of information as a vehicle for maintaining coalitions, it creates an artificial context for the building of a coalition and through that context, allows individuals to explore and express the meaning they see in the information they encounter.
There are many things that fascinate me about this latter approach. Not least, because it highlights the importance of togetherness and convivial experience in the identification of meaning and the coordination of action. It also strikes me as interesting that such an artifical activity context looks remarkably like Beer's Syntegrity (see http://www.syntegration.com/). Integrating such approaches with a deep understanding of information, and the political games that are played in institutions produces a fascinating practical approach which addresses some of the problems that institutions face in a rapidly moving information environment.
No comments:
Post a Comment