Posts Tagged ‘contrarian stand’

I have been branding myself as a contrarion over at edFutures in what I see as a fundamental issue, not mere semantics, with how we started the course. All started from the OECD starter framework and the edFutures week 1 discussion that I attempted to initiate on not using trends as our starting point if we are to analytically propose a framework for alternate education futures. Here are some more contrarian views :). George Siemens defines trends as a pattern of change over time.

Is a trend a pattern? Lets take for example, the trend towards greater use of social networking tools by the young. Typically, statistical analyses will capture the increasing numbers – of people, networks/groups, shares etc to indicate a trend base on tracking the user.

Network analysis and other statistical analyses may reveal patterns in the phenomenon of social networking – whether temporal or not. Starts out as a fad or a keep up with the Joneses and over time emerges as a integral component of our digital life. There is a pattern of use there. Then perhaps there are patterns of network formation and sharing preferences. Some users are aggressive sharers and some are private (the social network encourages and favours aggressive networkers). There may be patterns of induction (one boy in the class gets on FB, everyone follows…isn’t that a cool thing?).

You may get a wide distribution between networks/collaborations that are ad-hoc to those that are sustained across many years at a high level of participation and commitment (like a Community of Practice). Different analyses may bring up evidence of complexity, emergence, adaptiveness and chaotic behaviour.These patterns reflect the underlying states of the players – people, locations, technologies, cultures etc. The patterns are empirically observable, maybe even predictable to some degree – like some predictions about how groups form and work gets done (Cogs model etc). So we are essentially talking about different classes or categories of patterns – induction/creation, use, management, destruction, expansion etc. – that describe the phenomena.

A trend, however, indicates motion in data constituted by the values of two or more variables. A trend tries to describe relationships between variables. A trend could also try to describe relationships among sets of variables. The trend is not the variables themselves but an indication of how well (or not) they relate to each other. Trends are used to describe these relationships in  the context of uncertainty (imperfect correlation) and often raises questions that need to be answered through analytical models. Perfect correlation occurs when there is no uncertainty in the empirically observed relationships. Trends can be temporal or not.

Let us think of a phenomenon as a set of states – each state being a definable set of conditions. Patterns could be extracted from the phenomenon based on generalizability (if there is such a word) of a subset of those conditions. E.g. an induction pattern could postulate that the adoption of facebook is an exponential function if it is a local neighbourhood, but a linear function if a teacher in a class of 50 demands everyone sign up as part of the course. One of the facets of a pattern is that it is discernible across a wide range of situations, enough for some analytical or scientific models to be usefully applied to studying it.

Various phenomena may be correlated or networked. Sets of phenomena may exhibit complexity or emergence or adaptiveness or chaos or increasing returns.

Then let us think about change as change in states of the phenomenon. So patterns that are changing or are in motion would be described by, say, a first order differential of some or all of those subset of conditions that define the state. So we are not looking to describe how the conditions correlate, but rather how the changes in the conditions correlate – e.g. how does the linear function in the example change if there are influential rebels in the classroom who are against social networking. The induction pattern that would emerge is a change over the induction pattern that was originally observed or modeled – maybe a variant or maybe a completely new type of pattern. Patterns can change and evolve.

Next, let us look at it from another perspective. Can we discern patterns in changes of state? – yes, we can if there are patterns of change to be found. Perfect correlation: Heat water and at a particular temperature, it will start bubbling; sustain the heat and it will evaporate. Imperfect correlation: (Maybe :)) start CCK08 and ‘n’ variant open courses will follow; complete those successfully and they will spawn ‘N’ other variants of open courses. Those patterns of change may occur over time or over other parameters.

Could you find a trend in how these patterns change or the sequence/cycle of patterns? Maybe. But, is the trend the pattern itself? No.

A phenomenon is not a trend. A state is not a trend. A pattern is not a trend. A trend is not a pattern. A trend is not a pattern of change. Trends can indicate, describe, inform, influence our thinking of futures, but cannot act as the basis for an analytical model that we are seeking to develop.

Read Full Post »

%d bloggers like this: