Non-words for Test

Moderators: Debbie Hepplewhite, maizie, Lesley Drake, Susan Godsland

Elizabeth
Posts: 993
Joined: Sat Nov 20, 2004 8:47 pm

Non-words for Test

Post by Elizabeth » Mon Nov 22, 2010 9:40 pm

This is an important topic in its own right, so I'm starting a new thread. If follows the messages posted by Mona, Jenny and Maggie under OFSTED support for phonics on the third page and is linked to the one from Susan about the SP Reading Test.

For formative testing, asking children to read non-words is useful, as it helps their teachers to diagnose gaps in knowledge of grapheme-phoneme correspondences and difficulties with blending.

Six reasons not to use non-words for summative testing:

1. It would be unfair to penalise children using logical, but unexpected pronunciations.

2. It would be too restrictive to use only words where alternative pronunciations are extremely unlikely.

3. It would be unrealistic to give teachers enough acceptable alternative pronunciations to account for all logical possibilities.

4. It could be used by those who do not want synthetic phonics as evidence that synthetic phonics does not help with reading for meaning. This is not logical, as the ability to decode unfamiliar words is essential to reading for meaning. However, the argument could be used to undermine synthetic phonics. It is happening already.

5. It could have a negative influence on teaching. Here are some possible outcomes:

a) To avoid the confusion of possible alternative pronunciations only very straightforward words are used in the test (e.g. dop, mip, skib, hest). Then teachers stop teaching in a systematic way after they have finished teaching a very simple code, leaving children with insufficient knowledge to read and spell well.

b) More complex words are used. Then teachers try to teach rules to help children read words with graphemes with alternative pronunciations. This is not a good idea for the following reasons:
• There are nearly always exceptions to rules and even the probabilities are complicated.
• It is not a good idea to try to teach 5 and 6 year olds lots of rules, even if they work. They are too young. I keep it simple and say to children, ‘Try one way; if it doesn’t make sense, try the other way.’ (e.g. Reading ‘kind’, saying it with a short ‘i’ sound doesn’t make sense, but with the long ‘i’ sound (/ie/) it does.)

c) In order to prepare children for the test, teachers and publishers emphasise the reading of non-words at the expense of reading real words. It may be that teaching non-words is, in itself, damaging – at least for spelling. Children may see a non-word, subconsciously remember it, and then use analogy to spell a real word incorrectly. (e.g. see non-word ‘cuk’ – an example from a commericial programme – and write ‘duk’ for ‘duck’)

6. It’s unnecessary. There are thousands of real words that could do the job as well as non-words.
Elizabeth

Elizabeth
Posts: 993
Joined: Sat Nov 20, 2004 8:47 pm

Re: Non-words for Test

Post by Elizabeth » Mon Nov 22, 2010 9:44 pm

No ‘correct’ pronunciations

Letter-sound correspondences in English are not fixed and so there cannot be a correct pronunciation for a non-word, only a likely one.

Children may be taught rules, but rules are usually unreliable and complicated, for example:
• The non-word ‘bry’ is like the short words ‘my’, ‘by’, ‘try’, so the most likely choice for its pronunciation is one which rhymes with these words. But ‘y’ is also often pronounced as in ‘yes’ and as in ‘happy’. Children could be taught that ‘y’ is pronounced /i-e/ at the end of short words. However, not only is this getting complicated for a six year old, but it is not consistent without further ‘rules’. For example, ‘any’ is a common short word where ‘y’ is pronounced /ee/.
• Should ‘swad’ be pronounced /swod/ as in ‘swap’? When the letter ‘w’ is followed by an ‘a’, the ‘a’ is usually pronounced /o/ (e.g. ‘swan’, ‘want’, ‘washing’). But, as usual, this cannot be called a rule, only a probability – ‘swam’ is not pronounced /swom/.

Children might logically use analogy to read:
• ‘lome’ as /lum/ because they pronounce ‘come’ as /kum/
• ‘fas’ as /foz/ because they pronounce ‘was’ as /woz/
• ‘chab’ as /kab/ because they pronounce ‘Chris’ as /kris/
• ‘bave’ as /bav/ because they pronounce ‘have’ as /hav/ or as /baev/ because they pronounce ‘brave’ as /braev/
• ‘youp’ as /yoop/ because they pronounce ‘you’ as /yoo/ (also ‘youth’, ‘troupe’, ‘route’)
• ‘kead’ as /ked/ because they pronounce ‘head’ as /hed/
• ‘gen’ as /gen/ because of ‘get’ or /jen/ because of ‘gem’

There are numerous possible examples – too many to cater for by giving administrators a list of acceptable alternatives.

Regional variations make this even more complicated. For example:
• Children might be expected to pronounce ‘dut’ to rhyme with ‘cut’. However, children from London might pronounce ‘dut’ as /doot/ (short /oo/ sound), because they know that ‘put’ is pronounced /poot/. Moreover, children from Manchester are likely to pronounce ‘dut’ as /doot/ (short /oo/ sound) because ‘u’ usually represents the same sound as ‘oo’ for them. (‘Luck’ and ‘look’ are pronounced in the same way by most people in Manchester.)
• Children from London might pronounce ‘dast’ and ‘darst’ in the same way, to rhyme with ‘fast’, but children from many other parts of England would probably pronounce these two words quite differently.
Elizabeth

Elizabeth
Posts: 993
Joined: Sat Nov 20, 2004 8:47 pm

Re: Non-words for Test

Post by Elizabeth » Mon Nov 22, 2010 9:59 pm

Here are a few real words that would do just as well as non-words for testing children's ability to decode without memorising whole words or using context. There are thousands of possibilities.

bud , caw , cog , coy , dab , dim , ebb , sip , eel , elk , eve , fee , fez , fib , fob , gum , gut , hag , hip , hob , imp , jag, jig , lag , lap , lob , mar , mid , nag , mop , oaf , paw , ply , pod , rem , rod , rug , set , sly , soy , spy , tad , tar , ups, wed , wig , yam , yap , yen , zap , zed , ail , ape , apt , ark , bap , bay , brisk , brink , brisk , clips , edit , faith , flick, flits , froth , flint , frill , glide , glint , grid , joint , joist , loin , pond , pride , prime , quill , quilt , quip , raid , rail , shift, skid , skill , skimp , slick , slip , smith , sniff , tribe , twig , twine , trips , adapt , alarm , blade , bland , blaze , boast, brand , brash , clash , coast , crag , crack , champ , enact , drape , draft , gland , glade , loft , void , blond , adorn, aloof , block , booth , broke , abort , bloom , brook , broth , close , croak , croft , drone , flock , crook , frost , froze, globe , gloom , floss , gloat , gloss , groom , loot , scoff , scorn , shoal , rook , scoop , stomp , swoop , slush , spool, stoop , shorn , stork , thorn , stool , sworn , thorn , breed , creek , creep , greet , sheen , sleet , tweed , tweet , florid, scorch , thorax , snort , bait , bribe , twine , shade , shank , sharp , shawl , slave , snarl , span , stack , stain , trait, twine , sigh , flight , fright , plight , thigh , dime , lime , wife , glide , grime , pride , swipe , tripe , chime , crime , mope, vote , hone , doze , mode , probe , choke , drone , quote , spoke , ford , sort , cord , cork , horn , barn , hark , part, yarn , charm , chart , darts , harsh , marsh , cart , larva , narch , dirt , ploy , cloy , ahoy , bray , sway , bail , bait , hail , lain , wail , coax , foam , moat , loan , roam , deed , deem , feed , heel , feep , leek , keen , reed , reef , reek , reel , seek , teen , weep , chirp , third , thirst , pert , verb , serf , burn , turf , surf , curd , lurch , spurt , blurb, spurn , bane , cane , dale , fake , fade , gape , gaze , lame , pale , rate , sake , sane , vane , tame , vale , wade , bait , leap , coil , snag , cutlery , stoat , plied , hake , abode
Last edited by Elizabeth on Tue Nov 23, 2010 10:16 am, edited 2 times in total.
Elizabeth

chew8
Posts: 4165
Joined: Sun Nov 02, 2003 6:26 pm

Re: Non-words for Test

Post by chew8 » Mon Nov 22, 2010 10:28 pm

Elizabeth wrote:For formative testing, asking children to read non-words is useful, as it helps their teachers to diagnose gaps in knowledge of grapheme-phoneme correspondences and difficulties with blending.
Surely, though, the purpose of this proposed Year 1 test is to provide a formative assessment which will help teachers to 'diagnose gaps in knowledge of grapheme-phoneme correspondences and difficuties with blending'. For example, under 'Purpose' in the consultation document we read 'Pupils who have not reached this standard at the end of Year 1 should receive appropriate support from their school to ensure they can improve their phonic decoding skills, and will then have the opportunity to retake the screening check'.

Please, RRFers, think very carefully indeed before trying to undermine this very serious attempt by the coalition government to ensure that teachers understand the importance of teaching decoding systematically in Reception and Y1 so that most children reach the target and those who don't are given the extra teaching they need to bring them up to scratch.

Jenny C.

User avatar
Debbie Hepplewhite
Administrator
Posts: 3653
Joined: Mon Aug 29, 2005 4:13 pm
Location: Berkshire
Contact:

Re: Non-words for Test

Post by Debbie Hepplewhite » Mon Nov 22, 2010 11:10 pm

I'm not at all against tests to see how children are doing, or how well the teachers are teaching - but, better still, go to the 'source' of children's teaching which is the methodology of the teaching and the teachers' ability to provide lessons well.

I realise that looking at results will give some kind of picture - but not a full enough picture.

As a country, we need to make sure that teachers are well-equipped with the knowledge and skills they need and the materials they need to teach well.

Whilesover the 'teaching methodology' continues to be choice, which it is, then it continues to be a lottery for what the children receive.

I know because my granddaughter is bringing home repetitive text books with words beyond her code knowledge and blending skills and she is displaying all the typical signs of trying to memorise the sentences and shutting down abruptly the second she reaches a new bit of text and hasn't gone through it at school.

How upsetting and galling is that after years of solid study and lobbying and helping others to get their children and students taught!

Do parents and grandparents wait, then, until the children are aged six to reveal their code knowledge and how well they can blend?

My granddaughter was already on the road to phonics through her nursery - and is getting some phonics in her reception class - but not backed up with the type of reading from books that she needs.

Mum's attitude is that she will teach her the phonics at home - sadly, not surprisingly, the methods of the teachers are the persuasive ones - because they ARE the teachers with the authority.

Why not test the teachers themselves rather than just relying on a test of the children to persuade teachers to teach phonics?

Elizabeth
Posts: 993
Joined: Sat Nov 20, 2004 8:47 pm

Re: Non-words for Test

Post by Elizabeth » Mon Nov 22, 2010 11:34 pm

Jenny, I take your concerns very seriously. I do not want to undermine this attempt by the government to ensure that teachers understand the importance of teaching decoding systematically in Reception and Y1. I support the idea of a word reading test for six year olds and want it to succeed. It is the use of non-words that I am concerned about.

I am afraid that the use of non-words will provide ammunition for a huge backlash. UKLA has already published a statement “which gives 5 clear reasons to show why a ‘non-word reading test’ would be unproductive”. I have listened to literacy co-ordinators, headteachers and advisers who, without knowing what I think, tell me they would not approve of a test using non-words. They appear to me to be people who support the teaching of phonics, but are influenced by the beliefs of whole language advocates. Although it is difficult to reason with Michael Rosen, Henrietta Dombey and UKLA about the merits of synthetic phonics, there are lots of influential people who might be a bit concerned about a word reading test and wonder if it is a good idea or not, but will be tipped to supporting the anti by a test that includes non-words.

There is a big difference between
1) the use of non-words by a class teacher sitting with a child in the class who has reading difficulties and
2) a national test “to inform parents and generate national standards” (consultation document).
The first can be administered and interpreted in whatever way the teacher finds helpful and not reported to anyone else. The second must be as “consistent as possible to ensure that the results are a reliable indicator of national standards. This means that there must be defined rules in place for all aspects of the assessment administration.” (consultation document) As soon as there is an attempt to make it consistent, there are the problems caused by our inconsistent written language. As soon as it is to be reported to parents and others, there are the dangers of it causing an over-use of non-words in teaching.

There is no need for non-words, so why use them when there are so many difficulties with their use?
Elizabeth

g.carter
Posts: 1859
Joined: Wed Nov 05, 2003 7:41 pm

Re: Non-words for Test

Post by g.carter » Tue Nov 23, 2010 8:41 am

Gove, Gibb et al must be banging their heads in despair.

Elizabeth reflects the concerns and criticisms of Whole Language followers and suggests only a very limited
non-word test would be effective ie dop, mip, skib, hest.
But thousands of words could be generated that could include correspondences that have only one or two likely pronunciations. Take 'oi' and 'oy' for example: boip, boib, boick, boid, etc. coib, croib, cloib, droib, droic,dwoib, dwoick, and so on through the alphabet, including all 20+ consonant blends.
Section 3 could include correspondences such as 'ow' with footnote to state that either /oe/ or/ow/ pronunciation is acceptable - strowp = either /stroep/ or /strowp/.
This is basically a good test and we should be able to support the government's attempt to clarify the importance of knowing a transparent code and the ability to blend sounds to make a word or legitimate pseudo-word . Whole language, multi-strategy taught children will not be in a position to do the test: even if 'legitimate' words were used these would not be considered to be fair by the Whole Language proponents because they would advantage middle class children with their vastly larger vocabulary.

What we need to do, is to concentrate on those 20% who are already struggling after six weeks of teaching and provide performance indicators for all programmes - otherwise we're blowing in the wind.

Of course teacher training is vital - but we cannot wait until that is in place and we should be very supportive of the government's efforts.

chew8
Posts: 4165
Joined: Sun Nov 02, 2003 6:26 pm

Re: Non-words for Test

Post by chew8 » Tue Nov 23, 2010 9:21 am

Elizabeth wrote:There is a big difference between
1) the use of non-words by a class teacher sitting with a child in the class who has reading difficulties and...
How many class teachers are at present doing this or are likely to start doing it if it's not required by national screening?
Elizabeth wrote:2) a national test “to inform parents and generate national standards” (consultation document).
The first can be administered and interpreted in whatever way the teacher finds helpful and not reported to anyone else. The second must be as “consistent as possible to ensure that the results are a reliable indicator of national standards. This means that there must be defined rules in place for all aspects of the assessment administration.” (consultation document) As soon as there is an attempt to make it consistent, there are the problems caused by our inconsistent written language. As soon as it is to be reported to parents and others, there are the dangers of it causing an over-use of non-words in teaching.
Surely it's possible to devise safeguards against these dangers.
Elizabeth wrote:There is no need for non-words, so why use them when there are so many difficulties with their use?
At the moment, people are theorising about the difficulties - we don't yet know for sure that they will turn out to be real in practice. If serious difficulties emerge during the trialling and piloting mentioned in the consultation document, I would hope that the government would think again.

I think there is a need for non-words if we want a stringent test of decoding.

As Geraldine says, Gove and Gibb must be banging their heads in despair if they know that not only the whole-language people but also some (not all) synthetic phonics advocates are opposing this aspect of their proposal.

Jenny C.

User avatar
Debbie Hepplewhite
Administrator
Posts: 3653
Joined: Mon Aug 29, 2005 4:13 pm
Location: Berkshire
Contact:

Re: Non-words for Test

Post by Debbie Hepplewhite » Tue Nov 23, 2010 9:33 am

http://www.education.gov.uk/consultatio ... =no&menu=1

I'm flagging this government's consultation material here, and then I'll do so on a new thread with a heading about the consultation.

Elizabeth
Posts: 993
Joined: Sat Nov 20, 2004 8:47 pm

Re: Non-words for Test

Post by Elizabeth » Tue Nov 23, 2010 9:35 am

Just a quick correction for now. Geraldine writes,
Elizabeth ... suggests only a very limited non-word test would be effective ie dop, mip, skib, hest.
No, that's not what I meant.
Elizabeth

yvonne meyer
Posts: 736
Joined: Tue Oct 21, 2008 4:58 am
Location: Melbourne Australia

Re: Non-words for Test

Post by yvonne meyer » Tue Nov 23, 2010 9:59 am

This is not my area of expertise but I have been following a discussion on this issue between some Australian reading scientists. My understanding from what I've read is that they think the test should have non-words only, and no real words at all, on the grounds that children with zero phonics knowledge can still read words.

They suggest that a test of just 20 non-words, graded in difficulty, with the test discontinued after a certain number of errors, would be more suitable than a test of words and non-words.
"... This would solve the problem of administering 40 items to children who are unable to read, and are unable to respond correctly to most or all of the words presented. It would also make the tests much shorter to administer, and avoid the negative effects of giving children a task which is clearly beyond them."

Elizabeth
Posts: 993
Joined: Sat Nov 20, 2004 8:47 pm

Re: Non-words for Test

Post by Elizabeth » Tue Nov 23, 2010 10:43 am

Yvonne writes
My understanding from what I've read is that they think the test should have non-words only, and no real words at all, on the grounds that children with zero phonics knowledge can still read words.
I believe even able children with good vocabularies and good visual memories could not read all the real words I listed without using phonics if there were no context clues. If someone proved to me that I was wrong and they could, then perhaps those children wouldn't need synthetic phonics.

However, I am less concerned about assessment on a small scale. My concern is about a national test and all the implications of a national test that I've already written about.

For example, one of the non-words Geraldine used as an example was "dwoick". I thought there were no English words with the ending "oick", but then I found "hoick"; I think that's the only one. Children with good visual memories might read "dwoick" correctly and then start spelling words like “soak” and “week” as “soack” and “weeck”. If someone as knowledgeable about phonics as Geraldine can suggest a word with such an unlikely spelling, just think what could happen if less knowledgeable teachers and publishers all over the country decided to increase the teaching of non-words to help children do well in the test.
Elizabeth

JAC
Posts: 517
Joined: Tue Nov 15, 2005 1:51 am

Re: Non-words for Test

Post by JAC » Tue Nov 23, 2010 11:23 am

I greatly dislike the use of weird-looking non-words, but that is by the by!. A compromise(?) would be to generate lists of genuine parts of multi-syllable words , rather than completely novel and also 'odd' nonsense words.
Surely no-one could object to that?
eg tech (nology)
meth(od)
sec(tion)
med(ic)
doct(or)
att(end)
prin(cess)

Elizabeth
Posts: 993
Joined: Sat Nov 20, 2004 8:47 pm

Re: Non-words for Test

Post by Elizabeth » Tue Nov 23, 2010 11:36 am

But I would object.

"attend" has a double "t" because it is made of "at" and "tend". "at" on its own has only one "t".

There are no common words in English ending in "ec".

Would you expect the "ch" in "tech" to be pronounced /ch/ or /k/? "ech" as /ek/ at the end of a word is extremely unusual.
Elizabeth

JAC
Posts: 517
Joined: Tue Nov 15, 2005 1:51 am

Re: Non-words for Test

Post by JAC » Tue Nov 23, 2010 11:40 am

Elizabeth- these were just random examples to describe what I meant by parts of real words. I would hope that any test for 6 year olds would take into account their age and what they are expected to have been taught.

Post Reply

Who is online

Users browsing this forum: No registered users and 12 guests