Monday, December 8, 2014

Christensen on Data

Rotman's Martin Prosperity Institute's hosted a conference today on Knowledge Infrastructure. The conference featured many great speakers, including Clayton Christensen.

Christensen's big message was that we have to be careful about depending too much on data, particularly data collected by other people. Data does not represent a universal truth: somebody decides what data to collect and what not to collect. That choice in and of itself can influence or even obscure truth.

He strongly advised people go 'dumpster diving' for their own data, so that they choose the pertinent data that exactly pertains to the question they're trying to answer. "Using somebody else's data won't lead to the truth". He argued that everyone who does important things creates their own data.

Direct data is similar to direct observation.  In my experience, observations have the greatest power when you observe them first hand. Just as physics tells us that electrical signals and waves attenuate over distance, so the power of observations attenuate for each step away from the original observer. Christensen was arguing that this applies also to data; it loses its power for explanation unless it's a dataset you create yourself to answer the specific question you have in mind.

Christensen also described the stages involved in creating good theory as a summary of some things he was currently thinking about. Clearly not an off-the-shelf presentation, Christensen said it was the first time for this presentation and that his presentations were crummy the first four times! Not crummy, for sure, but lacking the luminous clarity and inevitability of logical flow of the previous presentations I've seen.

I was saddened to hear that Clay is currently undergoing chemotherapy. He's suffered so many medical issues - heart attack, cancer, stroke, diabetes - but his cancer had been in remission.

No comments: