(Journal of the Simplified Spelling Society 1992/2 p20, later designated J13)
[Also on this page: English and Italian, German and English.]
[David Moseley: see Journal articles.]

Three Reports on Recent Research.


David V Moseley, 'How Lack of Confidence in Spelling Affects Children's Written Expression'.
In Educational Psychology in Practice, April 1989, pp.42-46.

Dr Moseley is Reader in Applied Psychology at the University of Newcastle, England and co-author of the best-selling.ACE (=Aurally Coded English) Spelling Dictionary.
Abstract by Chris Upward and David Moseley.


Correct spelling is a widely desired educational goal, yet children's actual performance and attitudes in English are not well attuned to this expectation. British teachers have recently faced conflicting advice, with priority sometimes given to technical accuracy (eg correct spelling), but sometimes to content and style instead. It is claimed, on the one hand, that children's writing is inhibited by spelling difficulty and would blossom if correct spelling had lower priority, but, on the other, that correct spelling is a prerequisite for effective communication.

Past studies.

In one instance, it was observed that if a child had used a wider vocabulary, her spelling accuracy would have declined. A broader study showed a representative sample of twenty 15-year-olds averaging one misspelling in 60 words, with 6 pupils misspelling at least one word in 20, and 3 misspelling at least one in 10. One might suppose such pupils avoid using more sophisticated vocabulary with more difficult spellings.

A survey of 1254 pupils aged 8-9 showed the size of their spoken vocabulary correlated poorly with spelling accuracy. (Using high frequency vocabulary to test spelling obscures such correlations anyway.) It is therefore probable that poor spelling increasingly inhibits such pupils' powers of written expression at more advanced levels. Furthermore, misspelling of basic vocabulary is likely to prejudice examiners against the content and expression of such writing too. Moseley's 1989 study further emphasizes the dangers of deducing linguistic impoverishment from limited writing powers.

New evidence.

The study shows poor spellers using 35% more regularly spelt words than good spellers, and pupils diagnosed as dyslexic using 47% more. The poor spellers and dyslexics also used significantly more short words than the good spellers. Calculations were based on the proportion of words written which occurred in a list of 500 commonly used words.

Evidence was next sought which might show poor spellers avoiding irregularly spelt words, as well as less common words. Using scripts from 1,250 New Zealand pupils aged 7-13, a list was compiled of 51 words which occurred both among the 300 most frequently used and among the 300 most frequently misspelt. A second list was then compiled of 51 words of equal length also occurring among the 300 most frequently used, but not found amongst the 300 most often misspelt. It was predicted that poor spellers and dyslexics would use fewer of the 51 harder spellings than of the 51 equally common easier spellings.

The prediction was fulfilled when scripts by 20 British pupils aged 15 were examined. No difference in frequency of usage between easier and harder spellings was found amongst the 10 better spellers, but the 10 poorer spellers were found to prefer the easier spellings by a factor of almost 4:1. Similar results were obtained from dyslexics. This suggests that difficult spellings restrict pupils powers of expression, though less so in the case of younger New Zealanders.

Certain pupils had informed Moseley that they deliberately repeated words with safe spellings, rather than risk error by varying their vocabulary more widely. This was confirmed by a sophisticated comparison of type-token ratios and error frequency; this showed a moderate correlation. Poor spellers made more errors, wrote less, used shorter words and a more limited vocabulary, and repeated themselves more often.

Poor spellers were seen to score lower marks than any of the above limitations by themselves would produce, so suggesting such weaknesses had a cumulative effect. It is not clear what the best policy would be in these circumstances: should freer spelling be encouraged, in the hope of releasing more of the writers' innate powers of expression? In what circumstances should misspelling be penalized? More research and new ideas are needed.

One thing however is clear: the above results are all grist to the mill of spelling reform.

[Journal of the Simplified Spelling Society, 1992/2 pp21,22 later designated J13]
[Gwenllian Thorstad: see Journals, Newsletters.]

Three Reports on Recent Research.


Gwenllian Thorstad. 'The effect of orthography on the acquisition of literacy skills'.
In British Journal of Psychology (1991), 82, 527-537.
Dr Thorstad was formerly tutor in Educational Psychology at the Tavistock Clinic, London, and now advises the Simplified Spelling Society on research. On 25 April 1992 she gave an address to the Society, reporting and inviting discussion on her research findings. The following summary has been adapted from a press-release announcing her address to the Society, originally drafted by Leo Chapman, the Society's Vice-Chairman and Public Relations Officer.

Literacy problems in English.

All standardized reading and spelling tests in the UK indicate that it takes 10 years for English speaking children to be able to read and spell most words, while it is said that Italian pupils take only one year in school to achieve the same standard of literacy. Therefore there is no need for these types of test in Italy or carefully graded reading books, and there is little provision for remedial teaching.

The cause lies in the irregular spelling of English words, such as daughter and enough, which defeats many British people, of whom at least one third are not good spellers, 15% are semi-literate and 5% are illiterate.

While it is understandable that most people who can read and spell correctly do not want the spelling changed, they do not realize the price that is paid by many for this conservatism. For those, severe educational and social consequences can ruin their lives. It takes away all pleasure in school, often prevents them from following a chosen career, can be a cause of truanting and later delinquency. This is because at least 5-15% cannot memorize words from their appearance, however many times they see them, but need to build them up from the sounds.

British children have less time for their other studies, because they have to spend so long learning literacy skills in English. Some children with a Specific Learning Difficulty (Dyslexia) can never learn by the 'look-say' method and need specialized phonological remedial teaching.

Compared with Italian.

This is scarcely necessary in Italy because every lesson automatically becomes a remedial one. Even two Italian children, who apparently had a severe Specific Learning Difficulty at 6.5 years, had made remarkable progress when followed up a year later. As a result, Italian schools do not have to provide remedial teachers, nor do the text books have to be carefully monitored for their level of reading difficulty.

In Italian every letter is pronounced except <h> which is always silent. In contrast, the English alphabet has only one consonant, <v>, which has one sound, cannot be produced by other combinations of letters and is never silent. The five vowels have some 48 different sounds. There are also many homophones, which are words that sound the same but are spelled differently such as led and lead, and homographs, which are words that are spelled the same but pronounced differently according to their meaning, such as tear (in crying) and tear (rip).

However, most difficult of all are the small, common words which are totally irregular, such as was and to. Of the 100 most common words used in children's books, only 33 are regular. This is the greatest hurdle of all for dyslexic children.

English and Italian learners.

Dr Thorstad's study was done with 95 English children learning English traditional orthography (t.o.) and 70 Italian children, both groups aged from six to 11 years. There were also 33 English children learning the initial teaching alphabet (i.t.a.), aged from six to seven years. The English t.o. and Italian children were divided into three age groups of the same average age and ability, the youngest age group having an average age of 6.5 years like the English i.t.a. group. They attended small schools where the teaching was good. They came from stable homes with caring parents, according to the teachers.

The children were tested with a passage of 56 words taken from an article in an Italian journal about making cement in the Arctic Circle and translated for the English children. It was deliberately selected so that the subject matter and vocabulary would be unfamiliar to them, so that they would have to depend on looking closely at the letters in a word when reading, or analysing the sound of the word when spelling.

At 6.5 years, the English t.o. children read an average of 3l words correctly, the English i.t.a. children read an average of 50 words correctly, and after only six months in school the Italian children read an average of 44 words correctly.

But in spelling, the results of the English t.o. children were much worse. They could only spell correctly an average of 13 words, the i.t.a. children an average of 27 words, but the Italian children an average of 45 words. The English children could not attempt most of the words, while the Italian children made only very slight errors.

By 10.5 years the average of the English t.o. children in reading was 54 words, while the Italian average was 55 words. But in spelling the English t.o. children only had an average of 44 correct, while the Italian children had an average of 55 correct. The passage contained a considerable number of easy words like the and of.

Difficult words compared.

In order to bring out the difference more clearly the eight most difficult words were compared. These eight words, similar in both languages, were looked at separately: cement/cemento, correct/corretto, literally/leteralmente, perceptible/percettibile, permits/permette, preparing/preparano, special/speciale and thermometer/termometro. The English t.o. children only read an average of one of these words at 6.5 years and seven words at 10.5 years, while the Italian children could read an average of five words at 6.5 years and nearly 8 words at 10.5 years.

But in spelling, the younger English t.o. children had no words correct and only two at 10 years, while the ltalian children had an average of five correct at 6.5 years and nearly eight at 10.5 years.

The differences actually were far greater in terms of the number of errors in each word, for the Italian children often had only one slight error in stress in pronunciation in reading because they did not know the word, whereas the English children could often only say the sound of the first letter.

The Italian children used a systematic phonological approach in reading unknown words, while the English children rushed through the reading sounding out the first letter or two in a word and then just guessing. Thermometer was read as their mother by the seven-year-olds and the motor or the monster by the eight-year-olds.

There was a marked difference between the assurance with which the Italian and English i.t.a. children tackled these long, unknown words and the distress expressed by the English t.o. children.

[Journal of the Simplified Spelling Society 1992/2 p22, later designated J13]
[Chris Upward: see Journals, Newsletters, Leaflets, Media, Pamflet 15, Book, Papers.]

Three Reports on Recent Research.


Christopher Upward 'Is traditionl english spelng mor dificlt than jermn?'
In Journal of Research in Reading, 1992/2, pp82-94.
Chris Upward is Senior Lecturer in German at Aston University (Birmingham, UK), Editor-in-Chief of the Simplified Spelling Society, and Chairman of the Society's Cut Spelling Working Group. The abstract is written in Cut Spelling.

English spelng and its role in litracy teachng are curently undr intensiv debate in english-speakng cuntris. Yet ther is no consensus as to th most apropriat teachng methods, nor as to th importnce of conventionly corect spelng eithr as an aim or as an obstacl in th learnng process. It shud therfor be useful to compare th efect on standrds of litracy of th iregulr spelng of english with that of mor regulr spelng in othr languajs.

Such comparisns ar dificlt because of educationl variabls between cuntris. Th study atemts to overcom th dificlty by anlyzng errs made by english-speakng students of jermn riting in both languajs. Th presumtion was made that if two spelng systms ar of equal dificlty, riters wil make fewr errs in thei mothr tong than in a foren languaj, altho countrvailng factrs cannot be excluded.

Th errs found wer categrized by typ and probbl cause. Categris for both english and jermn wer: 1) silent letrs, 2) mispelt shwa, 3) dubld consnnts, 4) rong letrs representing corect sound, 5) intrference from othr languajs. A furthr five mislaneus categris wer establishd, mostly caractristic of one languaj only (eg apostrofes in english, <ß> in jermn). Th riters wer found to hav made 6.87 riting errs per 1,000 words overal in english, but only 3.87 in jermn, ie th foren languaj was ritn 1.78 times mor acuratly than th mothr tong.

Howevr it was considrd that even this result myt undrstate the gretr dificlty of english spelng, since th riters ar presumed to enjoy th advantaj of much longr and closer familiarity with english than jermn. A series of 3 closer analyses therfor atemtd to establish how far th errs cud be atributed specificly to dificltis of th two spelng systms, and how far extraneus factrs myt hav produced errs. Th first of these analyses removed those typicly metaorthograficl errs not involvng th choice of letrs (eg punctuation); th resultng calculation showd english to be 2.22 times mor dificlt than jermn. Th secnd analysis removed spelngs presumed to arise from intrference from foren languajs, aftr wich English apeard 2.36 times mor dificlt. Th third analysis removed al othr errs (typicly 'slips') not obviusly atributebl to featurs of th spelng systm concernd. At this point, wher it is claimd that th truest comparisn of dificlty between th two languajs can be mesurd, the gretr dificlty of english sord dramaticly, to a factr of 6.98.

Th categris of err wer then furthr examnd. It was found that 63% of th errs made in english wer asociated with redundnt letrs, ocurng in th thre categris 1) silent letrs, 2) mispelng of shwa befor <l, m, n, r>, and 3) dubld consnnts. In conclusion, these findngs wer related to th Cut Spelng proposal for simplifyng english spelng mainly by removing such redundnt letrs, and it was claimd that th study had shown not merely th much gretr dificlty of english spelng compared with jermn, but th apropriatness of a Cut Spelng solution to th problm.

Th paper is itself also publishd in Cut Spelng.

Back to the top.