MIS-management data: Evidence based policy does not work
Why we need to collect less data from schools, not more
Published 14:24, 21 January 13
Just out of interest I searched for Big Data ‘trending’ in education. Why? Because last year I searched for Cloud and the year before it was Tablets. Fashion, but sure enough at the BETT computer show for education next week I can sign up for a seminar ... helping me to explore the possibilities of Big Data in schools.
I know, it’s all a bit silly but data-sets are indeed getting larger and ever larger in schools since as a matter of course we now collect and electronically record data on student attendance, performance, personal details (including behavioural and medical data) and much more.
A reductive approach to data collection means that more and more detail is added to a child’s data profile. Even so, even taking into account an MISculture which has an insatiable appetite to collect data the data-sets are still not so very large ... as in Big Data large. How can this be rectified? (The premise being that Big Data is a ‘goodthing’ because it’s trendy)
Big Data sets could happen in schools. The drivers would be:
- Automated data-collection,
- Aggregation of existing data-bases and
- The existence of a surplus of Cloud Servers
What shape these drivers would take is a matter for conjecture so here are three likely conjectures.
- Use RF tags for student registration and monitoring of movements in and out of the school (for child protection purposes naturally) - oodles of data could be generated.
- ‘Voluntarily consolidate’ schools’ data sets through the likes of CapitaONE
- G-Cloud...no shortage of underused New Labour data centres there, even though most are in Dublin
Apologies, it turns out my conjectures are actually already facts so that’s the HOW dealt with. Now for the WHY.
The ‘why’ has little to do with teachers and children in the classroom and everything to do with managers, local authorities and politicians. They love to collect pupil data of any kind.
Not real data you understand. Not data a teacher needs like a record of how little Jimmy is doing in his or her class (called a mark-book) but proper, manly data...Management Data!
What is Management Data and why is it big?
Management Data is aggregated data: e.g. total passes at levels of SATs, GCSE, GCE; average attendance; numbers of each ethnic minority; summer birthdays and free-school meals ... these will suffice to get us started.
Necessarily these data lose the fine detail of the individual from which they were originally captured, but now, really importantly, it is in a form useful to generate charts and presentations. From these in turn, generalisations can be drawn, extrapolations made and plans created.
It would be wrong (they feel) to draw up such futuristic plans from just one elementary class’ data. Obviously a bigger sample is needed. How big? ... Well you know the answer ... as big as possible. Big is good.
Now let’s take a reality check. This data is junk. Its aggregation produces averages, percentages and spreads ... pure data Valium which dampens both the signal and the noise.
Moreover, the data itself is no good. Much is made up, some even falsified and the validation/verification processes are patchy and often completely ineffective. If you get muddled with school stats you just ask the classroom teacher and duck. They don’t know their MIDIS from their SAT from their ALIS.
The end result, in the decade of the school MIS (Management Information Software), is an onanistic party of managers playing with their junk-data. It from this here that education strategy is planned.
Consider the following recent example to finish this blog.
Figures recently have shown that children in Local Authority A will not do as well as others in Local Authority B notwithstanding data-matching for demographics. As a result Ofsted the government watchdog is barrelling into the LAs to see what is going wrong, starting with the worst, Derby. Most poor performers are in the Midlands and the North...so no surprise there then.
You get the assumptions here? The data is ‘incontrovertible’, the conclusions inevitable:
Local Authorities ‘run’ schools in these areas. Local Authorities are the managers and so they must be managing wrongly.
It’s obvious. Only it’s not. Anyone who has visited the lower performing schools in the lower performing areas may be struck by two things. These are; how poor is the school leadership and how weak are the teachers. Neither have anything to do with the Local Authorities who did not employ them in the first place.
Even the Financial Times has analysed these data and agree that indeed the performance gaps are really there and that there is a massive North-South divide. So If I was a Senior Manager I would conclude that since the data is derived by Local Authority area then the reason for the discrepancy (statistically very significant don’t forget) is because:
- Some Local Authorities are more competent than others (Ofsted’s view)
- The leadership and management in schools in these areas are poor (the Unions’ view)
- The teachers are rubbish as no-one want to teach in (say) Derby (the kids’ view)
- The children are retarded ‘up North’ due to deprivation junk food and eating too many ferrets (surely no-one’s view).
The point is, not only is the data dodgy, (we teachers know just how dodgy it is) you can draw a myriad of self-serving conclusions. It follows that the emergence of Big Data in school will facilitate more junk data collection and a greater range of cherry-picked (to suit your prejudice or political agenda) conclusions ... help us please!
Let’s just stop now, forget evidence-based policy making, it’s already making a mockery of itself; collect less data and talk to teachers more.