John Lennon once memorably described the process of creativity as “doing bits that you join together”. And in many areas of activity there are a number of “bits” that one joins together, all of which have an effect on other “bits”. Putting that into #strokeandpose language, you could say that innovations play out in a complex socio-technical system, with a number of context dependent variables.
The systems mapping approach allows us to keep focusing on strategic issues in order to describe causes and effects – identifying key potential issues (and potential solutions) without needing to quantify these issues. Fundamentally we are visually describing cause-and-effect relationships between issues on a strategic level.
These maps are only useful where a context is described (and may be linked to the needs of a particular audience). Difference between systems dynamics and systems mapping is that the former deals with with quantified relationships. A map presents a mechanism but does not address the problem of measurement. That said, it must (says Adam Cooper) be possible to describe changes regarding the content of the boxes (even if this is just in terms of “more or less”).
So we (Simon Grant, Christian Voight, Mark Johnson [for the love of all that is holy, read this man’s blog!!], Adam Cooper, me) set to describing a student perspective on the use of learner analytics to enhance the quality of a course. Which causes us to spend much time defining our terms, and led to the realisation that we have two systems that have the name “analytics”, one being the personal-level feedback and intervention learners expect as a result of their learning activity, the other being the summation of data concerning this (and other issues) used as to inform decision making at a course or institutional level. Mark attempted to bring these back together using a management cybernetics approach, but other than this we didn’t get very far in actually sketching out what is a very complex set of interactions.
From looking at the work of other groups this was a common initial finding – the act of attempting to set out processes in this way seemed to be a great way of prompting critical conversations if not actually producing systems maps.
What we actually ended up doing (after a lot of interesting discussion) was coming up with a range of potential student “drivers” for engagement with higher education. As follows:
DRIVERS
1. Employment. (we split this into exchange value and use value of education, very much in the marxist tradition)
2. Pleasure. (we went for a split between joissance and plasir, after Lacan)
3. Esteem. (split between peer status [extrinsic] and self-worth [intrinsic])
1. Employment. (we split this into exchange value and use value of education, very much in the marxist tradition)
2. Pleasure. (we went for a split between joissance and plasir, after Lacan)
3. Esteem. (split between peer status [extrinsic] and self-worth [intrinsic])
(I’m sure that these drivers deserve further unpacking, this’ll maybe be done in a different blog post as it was slightly outside the scope of the session.)
Based on these needs, we agreed that a potential student would need investment-grade data on one (or more – depending on preference) of these in order to make a decision regarding whether, where and what to study, and we went ahead and listed a load of data sources you could probably fill in yourself. The interesting thing for me was the recurrence of the idea of risk management within this conception of education. Upping the financial stakes for students seems to have increased the effect of these risk calculations in decision making – this seems to have been a direct and possibly intended consequence of the Browne Review of HE Funding and everything that followed.
So… we can use a range of data to help manage risk, then? Well, part of the system mapping approach seems to be identifying precisely which data would be useful in making a particular decision. There was a lot of talk about the need for “filters” to support people in deciphering huge amounts of data in order to make a decision, and Systems Mapping could help us design these filters.
From an institutional perspective, it could also help make better use of the increasingly over-hyped suite of system-level learner analytics packages – meaning that data can be used reliably and defensibly for particular well-defined purposes. Mapping causality and relationships seems like a fundamental and almost trivial activity, but we found that the act of actually sitting down and doing it opened up a number of important questions about what people were actually trying to do with learning analytics and why.
As an absurd extrapolation of the data-driven approach, it is possible to argue for an endpoint which would result in the entire learning function of an educational institution actually being managed by a small shell script, responding to constantly updated metrics by instigating interventions. You can imagine a whole article about a Second Industrial Revolution where management functions become entirely automated (indeed in inventory management, futures trading and a few other areas this has already happened).
Leave a Reply