Quite often the conclusions we draw from a data set are biased towards what we want them to be, hence the Guardian arguing SATS are irrelevant in pupil testing – which they might well be since their primary purpose is as an evaluation of how good the teacher is.
But it is quite fun to take the underlying result and propose some alternatives that might produce the same outcome. Here some of the fundamentals from the source are these:
Teacher assessments of achievement are as reliable … as test scores at every stage of the educational experience. Teacher and test scores correlate strongly
Teacher assessments also predict additional variance in later exam performance and university enrolment
So alternative conclusions we might go for are thus:
Teachers can accurately predict how well their pupils will do in their SATS.
Which is, dear god, exactly what you would hope for. Teachers having responsibility for embedding those skills tested in SATS. Turn that around the existence of a clear set of tested basic skills causes teachers to understand whether their pupils have such skills, and attempt to improve any weaknesses.
SATS good you see.
The actual process of secondary education makes little improvement to students outcomes.
These assessments which can also be used to predict GCSE, A Level and university are all around the SATS taken at 11 years old. So the outcome of secondary education is predominantly determined by the input.
Would the Guardian like to discuss whether social mobility is low because genetics predisposes educational outcome, or if social mobility is low because there are so very few decent teachers?