Signals and use of data

Some parties would like education to be 'data driven', which is OK if one has a good understanding of data, and the complexity of the phenomena represented by the data.

But there are dangers in taking a simplistic approach to the use of data including

• the critical factors are often unknown and even unknowable (Deming)
• cause and effect are frequently separated in place and time
• we naturally give more attention to strong signals and than weak signals
• most recorded data relates only to strong signals
• clues for achieving progress are very often hidden in weak signals

For example, there are three useful signs of progress in relation to problematic student behaviour. In the usual order of progress, they are

• incidents are becoming less frequent
• recovery time is being shorter
• incidents are becoming less severe

All of these signals are weaker than specific incidents themselves. thus we continue to record incidents, and while we may observe incidents becoming less frequent, we are less inclined to actually record these kinds of observations. Of these three, most significant and the weakest indicator (signal) of progress is the second one: incidents are resolved more easily and quickly.

Unfortunately this sign of progress is frequently obscured by subsequent incidents, particularly because they may be as serious are earlier incidents. And this tends to imply 'no progress at all'. Worth considering? I think so.

Check out the Support Planner website for more on Signals.