Fair Test?

I know what I think about the standardised testing of tiny children, and it’s not pretty. However, if the ‘powers that be’ are determined to bring in a baseline test for children at the start of school, then it is important that they design a fair test. I have some questions, and I am not sure that anyone can actually answer these questions. Certainly, I am not convinced that these questions can ever be answered in a way that leads to a test which is reliable, accurate and ethical. One that is of value for children, parents and teachers, as well as for the system.

1. A baseline is most useful to practitioners and teachers when it is done at the point at which a child begins in a setting. That way it can inform the child’s learning right from the start. But if a baseline test is done at the start of the year, then the children will be different ages. For a child born in August, who starts in a Reception class at the age of four, a September born classmate is not only a year older, he or she is a quarter of their age again. It is easy to miss this point, particularly if you are not used to working with small children. Mathematically, comparing a four year old to a five year old is like comparing a twelve year old to a fifteen year old. It is unfair, unjust and makes for an unreliable test. (Actually, the difference is even more stark than this comparison suggests. During the period from birth to five years, children develop incredibly rapidly – this pace of development will never be matched at any other point in their lives.)

* Are we testing all children at the same age, or at the point at which they start school? If the test is at the start of their schooling, then we are not comparing like with like, because the children are different ages.

* Are we testing children for their baseline at the same chronological age? If we are, some of them may have been in school for almost an entire academic year, while others will have just started. Again, we are not comparing like with like.

2. Although many parents do not realise it, a child does not have to start school until the term following their fifth birthday. If they want, parents can keep their child at home, or in an early years setting, up to that point. Until they turn five, children are not in statutory schooling. For the purposes of government surveys, a four year old in Reception is classed as being in ‘childcare’.

* If the baseline test is done at the start of the year, it is being done on children who are not yet in statutory schooling. Some of them will not even be there yet – they will be in a different setting, or at home. Is it ethical to perform a standardised test to test a system of statutory schooling, on children who are in ‘childcare’?

* The standardised tests that take place later on in a child’s school career, and by which schools are held accountable, are all given at a specific time of the year, not at the point a child reaches a particular chronological age. If the baseline is done at the point of entry, later tests may be measuring children who have been in the system for different amounts of time. If the baseline is done when the child reaches five, and later tests are not done according to their age, then a comparison between the two is inaccurate.

3. In order to design a test, you need to figure out what you ‘value’: which bits of early development you wish to test or check. But early development is interwoven, and complex, with one aspect predicated on another being in place. It is also incredibly ‘slippery’ – it does not follow a linear pattern. Often children experience sudden ‘spurts’, times at which their development seems to speed up exponentially. The notion that children must all reach a pre-defined ‘level’ at the same point in their schooling is toxic and unhelpful.

* Some children come to certain aspects of learning later than others, but they are much stronger in other parts of their development. If the test only checks certain parts of a child’s development, they risk missing the overall picture and ‘down grading’ a child just because they do not fit the specified pattern.

* Children’s ability to behave, concentrate and focus is a key factor in how well they will settle into school and how effectively they will learn. If the test measures cognitive rather than social ability, then it misses out on the behaviour required for learning. A child can be very bright, but still struggle at school because of behaviour issues, particularly in the case of children with specific kinds of special educational needs. How do we ensure that the test takes account of the complexity of early child development?

4. By definition, a test tests what we ‘value’. Even where we do not intend it, the test becomes a ‘norm’ for children to reach. It sets out a benchmark of what we expect, a benchmark that the world will see. Some children will inevitably fall below the norm, while others will go beyond it. For those children with special educational needs or disabilities, a ‘normal’ pattern of development will be entirely different.

* Are we comfortable with the notion of creating a ‘norm’ – of telling parents that their child is, by definition, ‘abnormal’? Is it even ethical to do this?

* Does our test value the things that we as a society wish to value, or does it simply measure a child according to a set of educational benchmarks? How do we ensure that the two do not get mixed up in people’s minds?

* What are the ethical considerations of testing a child by a standardised measure, when that child has down’s syndrome, or autism, or asperger’s syndrome, or ADHD? How will this help the parents and children, who may already be struggling to ‘fit in’ to a society that finds it difficult to cope with difference?

* How do we avoid the tendency for the educational system to put labels on children, labels which may limit our perceptions of what a child can do? Why would we wish to put a label on a child, rather than celebrating what it means to be a human being, in all its glorious complexity?

* Should the test be ‘opt in’ rather than ‘opt out’, to allow parents to give informed consent? If the test is ‘opt in’, then surely the lack of data from some children means the data itself is unreliable?

5. Think about a newborn baby: she knows nothing about the world, she can do nothing (apart from eat, sleep and cry). Now think about a five year old. The difference is total. Mind blowing, really. The ‘Development Matters’ guidance from the original EYFS outlines the astonishing number of attributes, skills and abilities that a child develops in the first five years. There is a reason why the EYFS profile ends up being such a massive, all encompassing document, one that is all but impossible to reduce to numbers on a spreadsheet. But numbers on a spreadsheet are what our accountability system requires. The best assessment of young children is done through a process of observing, of getting to know the individual child and all his or her different stages of development.

* What does the ‘data’ we are looking for actually tell us?

* Why is it necessary or even ethical to perform such a reductive exercise on tiny children, one that puts us at risk of labelling them from the very start? If the answer to this question is ‘to test our system’ then might I politely suggest that we need to stop in our tracks and question our values.

* Finally, how on earth are we going to choose which parts of development we believe matter? If we use ‘Development Matters’ as a starting point, then how on earth do we reduce the complexity of early child development down to a few testable items. Or, to put it another way, how are we going to ensure that every child still matters?

This entry was posted in Children, Data, Testing. Bookmark the permalink.

7 Responses to Fair Test?

  1. jameswilding says:

    Not meaning to do anything other than confirm what we do, the CEM centre PIPS service assists us hugely in establishing a benchmark for our pupils in Reception, at the end of Reception, at the end of Year 2 and at the end of Year 6 to show development, progress and value-added. We find it really helpful and would not do with out the very useful data it gives us. http://www.cem.org/pips/pips


  2. julesdaulby says:

    I agree with you Sue – it is vital tests show reliability and validity. The number 1 rule of testing should be to do the minimal amount of tests required. The observation part of testing is equally important; what strategies are they using? Do they verbalise? Use analogy etc – all tests are a snapshot of that particularly moment and do not predict future performance. Testing for testing sake and with children so young concerns me greatly.

    What are they testing for?
    What will they do with that information?
    Will it be to measure progress of the child or to highlight perceived failings of a school/teachers.

    I don’t understand why teacher assessment is not adequate at this stage?


    • nemocracy says:

      Teacher assessment is the only measure that is adequate at this stage, unless one were also to add in information from parents regarding out of school progress in learning. Narrowing children’s achievement down to a set number of easily tested points does no justice to the hidden aspects of development which do not conform to a linear progression through the 3Rs. Isn’t this what is being valued in the idea of value-added? It leads to teaching to narrow aims dictated by narrow goals. This is what seems most dangerous, rather than concern at the effect of testing on children – it is the effect on teaching – children are very adaptable and testing can be administered with care so they don’t notice (although this is difficult with the PIPS style assessments mentioned by a contributor above). The PIPS, as I knew it, several years ago now, was designed to test children in the contrived situation of an on screen game and did not test what an early years environment should be designed to teach according to the EYFS. All sorts of well-intentioned efforts were made to build bridges and references between the 2 measures, but they were never well-adapted.

      Assessing against the EYFS is ponderous because there is at least an attempt here to show and measure the diverse. Baseline testing against the EYFS is hampered because of all the factors drawn out above by Sue re age, stage, time in school etc.

      Perhaps thought could be given to an audit of provision as a means of describing children’s experiences of school, asking questions about the development of children within environments and how this prepares them for later schooling. Attitude to school is of huge importance at this stage, and dare it be said – happiness at school. Test what we are doing to children rather than what is ‘inside’ them, because what we do is formative. Unfortunately it is also fuzzy and doesn’t go on a spreadsheet.


  3. Pingback: Now That’s Primary Blogging Volume Four | prawnseyeblog

  4. Gorillerof3b says:

    I worry about lack of validity i.e. that tests will not be measuring what they believe they are measuring, and that the results of these tests will force teachers to focus on showing progress against a narrow range of fairly arbitrary descriptors rather than educating children as they see fit. The curriculum is already so dominated and skewed by Sats, I feel that this will only add another layer of restriction. I just can’t face spending even more of my life gazing at data and working out how to further trim my broad and balanced curriculum in order to turn another set of boxes green.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.