The effects of training of morphological structureon spelling derived words by dyslexic adolescents

Styliani N. Tsesmeli1* and Philip H. K. Seymour21Department of Education, University of Aegean, Rhodes, Greece2Department of Psychology, University of Dundee, Dundee, UK

This exploratory study aimed to determine the effects of explicit instruction aboutmorphological structure on the spelling of derived words. A cross-sectional abilitylevel-design was employed in order to determine differences in response to instructionbetween dyslexic students aged 13 þ years and age-matched and spelling levelmatched control groups. The study was based on the word-pair paradigm (a base andderived word) and combined oral instruction with written materials. The interventionhad a substantial impact in enhancing the spelling of derivations by the dyslexicadolescents. Their gains were appropriate for their spelling level, stable two monthsafter the intervention, and generalized to untrained but analogous items in terms ofstructure and suffixation. Non-dyslexic younger participants matched in terms ofspelling level also showed training and generalisation effects of the same size as theirdyslexic counterparts, while the age-matched controls did not improve so muchbecause of ceiling effects. It is proposed that morphological awareness constitutes apositive asset for dyslexic adolescents that can be used efficiently to counterbalancetheir severe phonological deficiencies.

The importance to literacy of phonological skills is well established (Bryant & Bradley,

1985; Liberman, Shankweiler, & Liberman, 1989; Torgesen & Wagner, 1992) and

preschool phonological training has the potential to produce gains in reading

development (Byrne & Fielding-Barnsley, 1995; Lundberg, Frost, & Petersen, 1988) andto treat reading disability (Hatcher, Hulme, & Ellis, 1994). There is now a consensus

(Bryant, Nunes, & Bindman, 1999; Henry, 1993; Moats, 1998; Snowling, 2000) that other

aspects of literacy disability merit further investigation, especially the morphemic

patterns of language that could provide an effective complementary strategy for later

reading and spelling development.

Recent experimental evidence suggests that awareness of the internal structure of

words is linked to spelling development in a variety of alphabetic orthographies (Arnbak

& Elbro, 2000; Bryant et al., 1999; Bryant, Nunes, & Aidinis, 1999; Leong, 2000;

* Correspondence should be addressed to Dr. Styliani N. Tsesmeli, Department of Education, University of Aegean, 85100Rhodes, Greece (e-mail: stsesmeli@rhodes.aegean.gr).

BJP 470—7/10/2008—ANISH—309616

TheBritishPsychologicalSociety

1

British Journal of Psychology (2008), 00, 1–29

q 2008 The British Psychological Society

www.bpsjournals.co.uk

DOI:10.1348/000712608X371915

Senechal, Basque, & Leclaire, 2006), especially in its advanced stages. Scientists theorise

that the processing of morphological units aids the retrieval of orthographic patterns

during spelling. The role of morphological knowledge in learning to spell has been

emphasized when, for instance, a child should decide between two or more

phonetically acceptable spellings for the same sound in English (e.g. box and socks) or

when to spell silent morphemes marked only by morphology in French (e.g. il chant)(Bryant et al., 1999). Students with dyslexia are usually severely disabled in their spelling

performance in comparison with their age-mates, and, in many cases, with younger

children of the same reading ability. Besides, they continue to present serious spelling

difficulties late into childhood and even in adulthood (Bruck, 1992, 1993; Critchley,

1981; Deacon, Parrila, & Kirby, 2006; Hoien & Lundberg, 2000; Lefly & Pennington,

1991; Pennington et al., 1986, 1990). The difficulties appear to be most substantial in

the spelling of morphologically complex words, particularly derived forms. These

forms, due to their length and complexity, pose special difficulties for students withdyslexia, whose spelling performance reflects, apart from the established phonological

deficiencies, a lack of morphological skills appropriate for their age-level. A number of

studies have examined whether morphological knowledge is weaker in children with

identified spelling difficulties (Senechal & Kearnan, 2007). Evidence from spelling tasks

suggests that readers with literacy difficulties and/or with dyslexia exhibit poorer

morphological knowledge than their normal age peers (Carlisle, 1987; Leong, 1989;

Leong & Parkinson, 1995; Fowler & Liberman, 1995), fail to recognize that derivationally

related words involve additions to already familiar stems (Carlisle, 1987; Derwing,Smith, & Wiebe, 1995; Kemp, 2006) and their spellings are affected considerably by the

morphological transparency of the orthography (Leong, 1989; Senechal et al., 2006).

Given the effects of morphological knowledge to spelling, current suggestions

involve that systematic and sequential instruction of morphology is needed during the

elementary years of schooling (Senechal & Kearnan, 2007). However, intervention

studies in this field are very few and often confined to limited periods of training. Henry

(1988, 1993) conducted a wide ranging study with normal readers in grades 3–5. The

control group received traditional basal instruction while the experimental groups weregiven additional training (see Henry, 1990) in phonological and morphological skills,

including compounding and the identification of word structure according to origin

(Anglo-Saxon, Latin or Greek layers). Significant treatment effects occurred for most of

the measures in the pre- and post tests. In a further study of primary school dyslexic

students (Henry, 1993), the intervention programme enabled learning-disabled students

to make significantly higher gains on the morphological subtests than on the

phonological subtests. The gain in Spelling was more marked for upper grade students

than lower grade students, especially on roots and the prefixes/suffixes/syllablescategory. Although this study incorporated the teaching of morphological patterns to

normal and dyslexic individuals, the lack of appropriate control groups and the

intermixture of phonological and morphological subtests in the design make it difficult

to reach conclusions about the effects of the morphological factor on spelling

performance.

More recently, Nunes, Bryant, and Olsson (2003) attempted to distinguish effects of

phonological and morphological training to 7- and 8-year-old children’s reading and

spelling performance, with the inclusion of appropriate controls. The intervention wasconducted in small groups of four to eight children in 12 weekly sessions which

consisted of group games, the aim being to promote explicit understanding either

of morphological or phonological rules. Children were allocated in five groups:

2 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

(a) Morphological training only (N ¼ 55), (b) Morphological training with writing

(N ¼ 55), (c) Phonological training only (N ¼ 55), (d) Phonological training with

writing (N ¼ 55) and (e) Control group (N ¼ 237). The morphological groups were

taught about word stems and grammatical categories in relation to inflectional affixes

and derivational affixes, while the phonological ones were taught mainly about long and

short vowels combined with a variety of operations (blending, classifications etc.).Results showed that only the groups that received an intervention where instruction

was combined with writing showed more progress in spelling than the control group,

and while there were no discernible effects on the use of phonologically based

conditional spelling rules, morphological intervention effects were significant in

spelling, especially on derivational suffixes.

An extensive training study of morphological awareness in dyslexic students was

conducted by Elbro and Arnbak (1996; Arnbak & Elbro, 2000). The experimental group

consisted of 33 students aged 10 and 12 years with severe reading and spellingdifficulties (at least 2 years below expected reading level). The control group comprised

27 reading disabled students matched in chronological age, gender and IQ. Both groups

participated in remedial education in phonological awareness, grapheme-phoneme

recoding, spelling to dictation, and oral and silent reading. The experimental group

received additional training in morphological awareness of compounding and

derivational and inflectional morphology1 for about 45 min per week over a period of

three months. The pre- and post tests included assessments of morphological awareness

(oral tests of morpheme subtraction, morphological analogy, and new word production)as well as various reading and spelling tasks. The experimental group made gains relative

to the controls in morphological awareness and the spelling of compounds and

derivations. Elbro and Arnbak (1996) concluded that the study warranted replication

‘in order to reveal to what extent morphological decomposition in spelling is a

compensatory strategy in dyslexia’ (p. 237). They also concluded that morphological

training might be most effective if it included written material: ‘Students should hear,

see and write the morphemes they are supposed to learn to read and write’ (p. 238).

This suggestion is in agreement with Hatcher et al. (1994) who found that phonologicaltraining was effective only when combined with explicit ‘linkage’ to orthographicQ1

structures and also with earlier similar findings by Bradley and Bryant (1983).

Following these suggestions, we undertook a small exploratory study of the

effectiveness of morphographic training on the spelling of complex words by dyslexic

adolescents. The study employed a cross-sectional spelling level-design (Backman,

Mamen, & Ferguson, 1984; Mamen, Ferguson, & Backman, 1986) as the most

appropriate method to investigate spelling disability. Performance was compared with

age matched (CA) and spelling level (SA) matched control groups, the aim being todetermine whether improvement by dyslexic individuals as an effect of morphological

training is associated with their chronological age or spelling ability.

Q1

The study was based on the word-pair paradigm (a base and derived word) which

was first introduced by Derwing (1976) as a way of evaluating the word relatedness in

1 Inflectional morphology deals with the selection of the particular form or variant of a word that is appropriate to the syntacticcontext of a sentence or utterance (Matthews, 1991). Inflectional morphemes do not change the word-class or the meaning ofa word (e.g. book-books), such as the derivations. By contrast, derivational morphology refers to the creation of new lexicalitems by prefixation and/or suffixation of a single root (e.g. walk-walker). Compounding is another way to create new words. Itrefers to the process of combining two bases together which can be rare roots or affixed words (e.g. teapot, kind-hearted )(Katamba, 1994).

Morphological spelling training and dyslexia 3

BJP 470—7/10/2008—ANISH—309616

terms of meaning. In this study, each pair consisted of a base and a derived form and was

used as an index of the application of morphological strategies in spelling. It was

previously found (Carlisle, 1987; Tsesmeli & Seymour, 2006) that disabled readers and

younger children spelled derived words significantly worse than their bases and tended

to perceive the items as two different words (e.g. sun-sany instead of sun-sunny),

suggesting a failure to deduce unfamiliar words from more familiar ones. Kemp (2006)also found that transfer of spelling effects was no better than chance when the child

did not know how to spell the base word, but was over 70% when the base form

was known.

Training combined oral instruction with written materials, the aim being to train

students in the morphological structure of derived words and to make explicit the links

between morphological and orthographic structure. The intervention was targeted at

derived words since the derivation process is strongly linked to the creativity of the

language through the formation of new words. It constitutes a very central process forthe expansion of word knowledge from known to unknown items (Katamba, 1993). In

addition, derived words are typically long, low in frequency and abstract in meaning

(Nagy & Anderson, 1984), and create significant difficulties in spelling (Carlisle, 1987;

Kemp, 2006; Tsesmeli & Seymour, 2006).

In a deep orthography such as English the achievement of full competence in

spelling requires the coordination of a number of distinct categories of knowledge.

These include: (i) a grasp of the phonography of the language, encompassing the

alphabetic principle by which the phonemes of speech may be represented by lettersand letter groups; (ii) a word-specific lexigraphic memory for the exact letter choices

conventionally employed to identify particular words or free morphemes; and (iii) a

morphographic knowledge of conventions for representing bound morphemes

(prefixes, inflections, derivational suffixes) and combining them with word stems.

In some accounts (e.g. Frith, 1985; Seymour & Duncan, 2001) these categories of

knowledge are acquired cumulatively with phonography as the early foundation and

morphography as a later and more sophisticated development. According to this view,

compensatory morphological instruction might be most effective in the later stages ofdevelopment. In order to examine this point we tested the effects of morphographic

intervention on both normally developing and dyslexic individuals with widely differing

levels of orthographic development.

Morphographic instruction might help spelling by leading students towards the

realisation that a complex and difficult word, such as ‘darkness’, is in fact composed

of two simpler elements, ‘dark’ þ ‘ness’, which recur in other contexts, thus

reducing the number of distinct forms which need to be learned. However, in many

cases, the conventional spelling of derivations involves additional changes, forexample: ‘happiness’ ¼ ‘happy’ þ y ! i þ ‘ness’, ‘noisy’ ¼ ‘noise’ – e þ ‘y’. In these

instances, correct spelling requires learning of the standard forms for the free (base

word) and bound (suffix) morphemes and an understanding of the conditions

demanding implementation of certain orthographic transformations (delete ‘e’,

change ‘y’ to ‘i’, etc.). In order to separate these aspects we included derivations

involving simple attachment of a suffix to a base (no change items) and derivations

requiring a transformation (orthographic change items) in the intervention study.

We additionally wished to determine how far learning generalized from the specificitems used in instruction to other words having the same morphological structure,

and compared intervention effects for both taught and untaught items for this

purpose.

4 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

In particular, the main hypotheses for the intervention studies were the following: (i)

Findings (Carlisle, 1987; Elbro, 1990) suggest that the dyslexic group will perform below

the CA- and the SA-controls before the intervention. It is hypothesized that the CA-group

will gain less from the intervention than the other two groups, due to their high

performance before training. Gains of the dyslexic group are expected to be comparable

to those of younger children of the same reading level, (ii) Evidence posits that spellingof base words is better than derived words (Carlisle, 1987; Kemp, 2006; Moats, 1998).

It is anticipated that this effect will be present before the intervention, and that the

outcome of the training will be to increase performance on derived words in every

group, (iii) Spelling performance is affected by morphological complexity and students

perform better on transparent than on non-transparent items (Carlisle, 1987; Moats,

1998). It is hypothesized that each group will show better scores on No Change than on

Orthographic Change items on the pre-test, and that training will selectively enhance

performance on Orthographic Change items for each group, (iv) Training effects will beevaluated in relation to three word sets, trained, untrained but analogous, and untrained

and non-analogous words, anticipating that each group will perform better on

instructed words than on uninstructed words, (v) Generalisation effects will be further

examined hypothesizing that untrained but analogous items to trained ones in terms of

word-structure and suffixation will evoke transfer-of-learning effects (Freyd & Baron,

1982; Wysocki & Jenkins, 1987). Finally, (iv) durability of training effects will be explored

by carrying out a delayed post test, after two months of the completion of the study, only

with the dyslexic group, due to serious restrictions of time for the rest of theparticipants.

Method

ParticipantsThe participants followed mainstream schooling in secondary and primary education in

the Perth and Kinross educational district in Scotland. They were selected so as to formthree groups:

Dyslexic groupThis consisted of nine male individuals from the 2nd (N ¼ 5) and 3rd (N ¼ 4) years of

secondary education, mean ages 13.9 years (range: 13.4–14.5 years) and 14. 9 years(range: 14.11–15.3 years). They were suggested by the principal teacher of the Learning

Support Unit based on assessments (by educational psychologists) of intelligence, literacy

and cognitive function and after they had been issued with a statement of special

educational needs (DfEE, 1994). They attended a Learning Support Unit for between one

and ten sessions per week and most received educational help from Learning Support

Assistants. No participants with any sight, hearing or serious health problems were

included. Their deficiencies on basis of disruption of accuracy of reading and spelling

processes (British Psychological Society, 1999) were further verified by completing acognitive assessment battery (Seymour & Evans, 1993) that contained tasks of word and

non-word reading and spelling. Items from tasks were regular (e.g. cut), rule-based

(e.g. air) and irregular (e.g. who) words of graded difficulty stratified in terms of word

frequency and length to represent three levels of orthographic complexity. There were

large discrepancies between individual dyslexic scores and mean scores of the

Morphological spelling training and dyslexia 5

BJP 470—7/10/2008—ANISH—309616

CA-matched controls (by p , .001) on both word/non-word reading and spelling. Error

analysis showed that dyslexic individuals presented significantly higher rates ( p , .001)

than their classmates on word substitutions (e.g. brush instead of bunch), regularisations

(e.g. autom for autumn), and severe phonological distortions (e.g. pipint for puppet)

(Tsesmeli, 2002). Individual data are shown in Appendix I.

Chronological age (CA) control groupThey were secondary 2 and 3 students (N ¼ 14) with mean ages of 13.7 years (range: 13.4 to 14. 3 years) and 14. 9 years (range: 14.7 to 15.4 years). They were selected through

the English department of the school as having average performance in National Tests of

reading and spelling. None of them attended the Support Unit and the majority were

classmates of the dyslexic group.

Spelling age (SA)/reading age (RA) control group2

They were selected from secondary 1 and primary 4–7 classes to reflect the range of

reading and spelling ability of the members of the dyslexic group3 (N ¼ 23 for the

Adjective Study, N ¼ 22 for the Noun Study). Selection was based on class teachers’

identification of average readers who presented no particular difficulties in reading or

spelling relative to their age.All participants undertook a psychometric assessment consisting of Raven’s Standard

or Coloured Progressive Matrices4 (Raven, 1958; 1962) and the British Abilities Scale

(BAS) subtests of Word Reading, Spelling and the Digit Span (Elliot, 1992; Elliot, Murray,

& Pearson, 1983). Summary results for the three groups have been included in Table 1.

Following analysis of variance, post hoc (Tukey HSD) tests showed that the dyslexic

group did not differ from the CA-group in age or intellectual ability and was equivalent

to the SA-group in reading and spelling age. However, there was a difference in Digit

Span between the dyslexic group and both control groups ( p , .001).

THE INTERVENTION STUDY

Experimental stimuliThe items devised by the first author and included two lists of word pairs, the adjective

list (N ¼ 97 pairs) and the noun list (N ¼ 100 pairs). The derivation process is

2 Spelling Age (SA) control group is identical with Reading Age (RA) control group on this sample. For the sake of brevity only theterm of the SA-group is used throughout the paper. Spelling ages in each group were somewhat lower than their reading ages.Details for the dyslexic group are given in Appendix I. Spelling and reading ages are in parentheses for the CA-group(RA ¼ 14.20, SA ¼ 12.93) and for each sub-group of the SA-group (P4: RA ¼ 7.87, SA ¼ 7.24, P5: RA ¼ 8.85,SA ¼ 8.67, P6: RA ¼ 9.64, SA ¼ 9.26, P7: RA ¼ 11.85, SA ¼ 10.47, S1: RA ¼ 15.47, SA ¼ 10.87). The correlationbetween reading and spelling age was strong and significant in both dyslexic (r ¼ .777, p , .01) and control groups(r ¼ .840, p , .001).3 The variability in Spelling Age (SA) scores in the dyslexic group (see Appendix I) did not allow us to form a homogenousSA-group in terms of spelling ability but rather a ‘stratified’ sample. Thus, singles or dyads of dyslexic individuals of common SAmatched with a small group of similar SA of particular Grade, since it is not possible to choose pupils with normal spelling skillsof low SAs from high Grades and the reverse. For example, dyslexic pupils with SAs of 7.02, 9.01 or 10.10 matched withgroups of younger pupils but with normal spelling skills with means of 7.24 (Grade 4), 9.26 (Grade 6) and 10.47 (Grade 7) asappropriate. This led to a variation in SA within the SA-group.4 The Standard Progressive Matrices was used for the Dyslexic group, the CA-control group and the S1 of the SA-controlgroup, while the Coloured Progressive Matrices was applied to the P4 to P7 of the SA-control group.

6 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

applicable only to content words (i.e. nouns, adjectives, verbs and adverbs), thus

adjectives and nouns derivations were chosen as they are among the most

representative of derivational suffixes in English (Katamba, 1993). The selection of

derivational items was based on pilot studies undertaken to identify suffixes which were

particularly problematic for dyslexic students (e.g. -ous: 3.33%, -ance: 7.77%, -ful: 15%,

Tsesmeli (2002)). The derivational items in the adjective list contained five different

suffixes: (i) 20 items in –y (LUCK-LUCKY); (ii) 17 items in –ful (USE-USEFUL); (iii) 20

items in –ous (DANGER-DANGEROUS); (iv) 20 items in –ive (ACT-ACTIVE); and (v) 20items in –al (NATION-NATIONAL). The suffixes for the noun list were: (i) 20 items in –

ness (DARK-DARKNESS); (ii) 20 items in –ance (PERFORM-PERFORMANCE); (iii) 20

items in –ity (POPULAR-POPULARITY); (iv) 20 items in –ion (PROTECT-PROTECTION);

and (v) 20 items in –er (TEACH-TEACHER). The suffix types were crossed with two

levels of morphological complexity5: (i) NO CHANGE (N ¼ 50), where there is no

change between the base form and its derived counterpart as in STEM þ SUFFIX (e.g.

LUCK LUCK þ Y or DARK DARK þ NESS); and (ii) ORTHOGRAPHIC CHANGE

(N ¼ 47 for adjectives, N ¼ 50 for nouns), where there is an alteration in the base wordwhen the suffix is added as in STEM þ RULE þ SUFFIX (e.g. SUN þ Y¼ SUN þ N

þ Y; NOISE þ Y¼NOIS þ Y; HAPPY þ NESS ¼ HAPP þ I þ NESS). Despite their

inconsistency, these three derivational rules (i.e. consonant doubling, ‘magic –e’, and

‘change -y- to -i-’) are the most common rules in English that are applied to the majority

of items involving morphophonemic changes (see also below the Section Quality of

errors, and Tables 5 and 6 about this issue).Q2

For the purposes of the training study, each list was divided into three subsets of

items: (1) Trained items (N ¼ 40) were the word pairs used in the teaching programme.They included the suffixes –y, -ful, -ous and –ive for the Adjective Study and –ness, -ance,

-ity and –ion for the Noun Study, (2) Untrained 1 items (N ¼ 37 for adjectives, N ¼ 40

for nouns) similarly contained items ending in –y, -ful, -ous and –ive for the Adjective

Study and –ness, -ance, -ity and –ion for the Noun Study. These words were analogous to

trained items in terms of word structure and suffix-type (e.g. RUST-RUSTY analogous to

Table 1. Psychometric data for the groups

BAS

CA RA SA Raven’s PM Digits

Dyslexics 13.87 (0.62) 9.14 (2.06) 8.13 (1.10) 37.60 (4.90) 25.40 (22.6)CA-group 13.77 (0.71) 13.79 (0.62) 12.40 (1.13) 44.90 (6.91) 55.40 (26.9)SA-group 9.64 (1.75) 10.17 (2.11) 9.30 (1.78) 26.44 (9.73) 58.84 (32.3)

Chronological age (CA), reading age (RA), spelling age (SA), standard/coloured progressive matrices(raw scores) and BAS-digit recall (centiles).

5 Although a number of other changes could have been envisaged (e.g. phonological only, both phonological andorthographic, see Carlisle, 1987; Leong, 1989), orthographic change was the prevalent feature for these pair-itemsinvolving the most problematic derivational suffixes for dyslexic students. However, a few items involving phonologicalchanges (usually stress shifts: 23/197 or vowel alternations: 3/197) were unavoidable, due to scarcity of appropriate itemsin each category. These cases are noted in Appendices II & III and were reasonably well balanced between the No Changeand Orthographic Change categories. In addition, while phonological changes are shown to affect mostly oral production ofderived items (Carlisle, 1987; Fowler & Liberman, 1995; Leong, 1989), orthographic change appears to affect moreprominently their spelling (Tsesmeli, 2002).

Morphological spelling training and dyslexia 7

BJP 470—7/10/2008—ANISH—309616

LUCK-LUCKY or MAD-MADNESS to DARK-DARKNESS) but they did not receive any

direct instruction, and (3) Untrained 2 items (N ¼ 20) did not relate in any way to the

trained items and did not receive any instruction. They ended in –al (e.g. POST-POSTAL)

for the Adjective Study and in –er (e.g. TEACH-TEACHER) for the Noun Study.

A full listing of items along with frequency and letter-length details is given in

Appendices II and III.

General procedure of the studyThe general procedure of the study is outlined in Table 2. Each separate study included a

pre-test, a training programme and a post test. A delayed post test was given

approximately two months after the completion of the study only to the dyslexic group,due to serious restrictions of time for the rest of the participants. Each study lasted

approximately three months, the Adjective Study preceded the Noun Study and both

were completed during the last semester of a school year.

Assessments before and after the teaching programmeAll the items for each study (see Appendix II and III) were randomised to form the pre-,

post- and delayed post tests. Each test was given individually to the student by the first

author and lasted two sessions of about 40 min. Both base and derived words were

instructed to dictation as a pair (e.g. LUCK-LUCKY). The students had to write down the

spellings on three A4 sheets marked with two separate columns, placing the basespelling in the left column and the derivation in the right column. Although this

procedure favours immediate recognition of the stem–derivation relationship, contrast

to other practises in experimental literature (Derwing, 1976; Derwing et al., 1995), it

appears to evoke more sensitive results. Surprisingly to an experienced reader, we found

Table 2. General procedure for the intervention study

A. Pre-testing1. Pre-test of the Adjective Study 97 pairs randomised 2 sessions2. Pre-test of the Noun Study 100 pairs randomised 2 sessions

B. Training programme1. Adjective Study 40 pairs in a systematic order 8 sessions- Y (NC, OC) Luck-Lucky/Fun-Funny 2 sessions- FUL (NC, OC) Use-Useful/Beauty-Beautiful 2 sessions- OUS (NC,OC) Danger-Dangerous/Fame-Famous 2 sessions- IVE (NC,OC) Act-Active/Expense-Expensive 2 sessions

2. Noun Study 40 pairs in a systematic order 8 sessions- NESS (NC, OC) Dark-Darkness/Happy-Happiness 2 sessions- ANCE (NC, OC) Perform-Performance/Guide-Guidance 2 sessions- ITY (NC,OC) Personal-Personality/Active-Activity 2 sessions- ION (NC,OC) Act-Action/Educate-Education 2 sessions

C. Post-testing1. Post-test of the Adjective Study 97 pairs randomised 2 sessions2. Post-test of the Noun Study 100 pairs randomised 2 sessions

D. Delayed post-testing (only for dyslexic students)1. Delayed post-test of the Adjective Study 97 pairs randomised 2 sessions2. Delayed post-test of the Noun Study 100 pairs randomised 2 sessions

8 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

(Tsesmeli & Seymour, 2006) that our dyslexic students on different spelling tasks from

the present study presented significantly ( p , .001) higher inconsistent stem spelling

scores of misspelled pairs (44%) than CA- and RA-groups (5 and 27%, respectively)

showing that familial relationships of a word-pair are not so salient for the dyslexic

students as for normal readers of both groups.

Training programmeThe training programme was also implemented by the first author in individual sessions

of about 40 min for each student of the three groups. There were eight separate

Teaching units for every study, each involving instruction on five word pairs belongingto a particular suffix-type and morphological complexity condition. The No Change

items always preceded the Orthographic Change items. The sequence of instruction in

terms of suffix-type and condition was the same for all students (see Table 2).

The programme aimed to teach students in a step by step way the internal structure

of the words and how this related to their spelling. Each session had a sequential and

structural nature and was based on the active participation of the individual. More

specifically, instruction was targeted towards four main principles: (i) word structure -

every derived word was composed of a stem and a suffix; (ii) stem consistency – thebase and the stem of the derived word were spelled identically; (iii) suffix consistency –

the suffixes were spelled identically despite their different linguistic environments; and

(iv) derivational rules - in Orthographic Change items rules are applied in a systematic

way. The sessions were divided into three main phases, referred to as the Workshop, the

Discussion, and Practise in word-pair spelling (see Table 2). The Workshop based on

student’s own implicit understanding of word structure and can be viewed in Appendix

IV. The child was presented with a pair of soft cards and asked to analyse the word into

its constituent morphemes, to spell the stem, the suffix and finally to create the newword by combining the stem and the suffix appropriately. The Discussion phase aimed

to make the relationships between familial words more salient. To this end, the word

pairs printed on coloured cards where base words and the stems of derivations were

printed in blue, their suffixes in green and letters subject to change due to derivational

rules in yellow (e.g. HAPPINESS). All the coloured cards were laid together to facilitate a

discussion on the principles of stem and suffix consistency. Spelling practise followed

Discussion and aimed to make explicit how the knowledge of word structure links with

the spelling of the word pairs based on accuracy. This means that the child should writeeach word-pair at least 3 times correctly before proceeding to the next pair.

Scoring procedureThe spellings produced in the pre- and post tests (and in the delayed retention test) wereclassified as conventionally correct or as errors. Every accurately spelled word was

assigned 1 point and every misspelled word 0 points (e.g. wurmth instead of warmth).

Results

Table 3 reports the mean accuracy rates for the dyslexic group and for the two control

groups in the Adjective Study. Table 4 gives a comparable summary for the Noun Study.

The overall appearance of the data is very similar in the two studies. In the pre-tests

dyslexic performance was massively impaired relative to the results for the

Morphological spelling training and dyslexia 9

BJP 470—7/10/2008—ANISH—309616

Table

3.Theadjectivestudy

Basewords

Derivations

Nochange

Orthographicchange

Overall

Nochange

Orthographicchange

Overall

Dyslexics

pre-

47.55(18.99)

28.36(15.77)

37.95(17.13)

21.11(19.31)

13.94(11.70)

17.52(14.83)

post-

56.00(16.30)

35.93(19.16)

45.96(17.31)

49.55(18.70)

28.13(19.43)

38.84(18.91)

Gain

8.44

(8.64)

7.56(10.21)

8.01

(7.86)

28.44(12.31)

14.18(12.94)

21.31(10.84)

delayed

post-

57.11(16.64)

38.53(19.04)

48.11(17.48)

43.55(22.40)

25.05(22.66)

34.59(21.65)

retention

9.55

(8.41)

10.16

(9.67)

9.85

(7.97)

22.44(13.77)

11.11(13.44)

16.95(10.86)

SA-group

pre-

64.95(21.68)

50.69(23.78)

57.82(22.35)

48.17(26.45)

33.30(26.30)

40.73(25.68)

Post-

74.26(16.47)

61.51(22.10)

67.88(18.99)

70.78(18.90)

54.11(24.76)

62.44(21.42)

Gain

9.30

(9.54)

10.82(12.95)

10.04

(9.98)

22.60(17.98)

20.81(21.14)

21.70(17.96)

CA-group

pre-

94.00

(4.15)

85.56

(7.13)

89.78

(5.19)

86.14

(8.89)

70.06(17.67)

78.10(11.80)

Post-

95.28

(2.01)

90.57

(5.39)

92.92

(3.36)

93.00

(4.42)

85.25(10.20)

89.12

(6.46)

Gain

1.28

(3.81)

5.01

(4.54)

3.14

(3.18)

6.85

(7.83)

15.19(11.46)

11.02

(7.69)

Meanaccuracy

rate

(%)forbasewords/derivationsandNoChange/O

rthographicChange

item

s.

10 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

Table

4.Thenounstudy

Basewords

Derivations

Nochange

Orthographicchange

Overall

Nochange

Orthographicchange

Overall

Dyslexics

pre-

46.44(20.48)

28.66(15.87)

37.55(17.66)

33.77

(20.16)

21.11(18.16)

27.44(18.86)

Post-

60.66(21.79)

39.55(21.20)

50.11(21.20)

54.22

(23.37)

34.44(23.51)

45.83(23.00)

Gain

14.22

(9.35)

10.88(10.30)

12.55

(8.17)

20.44º(13.48)

13.33

(9.94)

16.88(11.04)

delayed

post-

54.44(20.85)

40.00(20.97)

47.22(20.63)

45.33

(24.85)

27.33(24.57)

47.22(20.63)

retention

8.00

(5.00)

11.33(11.48)

9.66

(6.12)

11.55

(11.73)

6.22(11.50)

19.88

(8.88)

SA-group

pre-

69.90(14.32)

55.09(21.38)

62.50(17.28)

57.00

(24.17)

38.72(25.61)

47.86(24.34)

Post-

80.63(15.21)

68.54(19.31)

74.59(16.85)

77.36

(20.92)

66.18(21.44)

71.77(20.57)

Gain

10.72

(7.64)

13.45(10.40)

12.09

(6.61)

20.36

(15.32)

20.81(21.14)

23.90(12.92)

CA-group

pre-

95.71

(4.06)

89.42

(7.97)

92.57

(5.58)

91.42

(5.73)

81.42(18.23)

86.42(11.26)

Post-

97.57

(2.95)

89.71

(7.31)

93.64

(4.73)

96.42

(3.34)

88.28

(8.03)

92.35

(5.37)

Gain

1.85

(2.76)

0.28

(3.22)

1.07

(2.30)

5.00

(3.98)

6.85(11.11)

5.92

(6.64)

Meanaccuracy

rate

(%)forbasewords/derivationsandNochange/O

rthographicChange

item

s.

Morphological spelling training and dyslexia 11

BJP 470—7/10/2008—ANISH—309616

Chronological Age control, which were close to ceiling, and fell below the level of the

Spelling Age control. Derived words were spelled less accurately than base words, and

accuracy was lower for Change than for No Change derivations.

The effectiveness of the intervention was initially evaluated in terms of the

improvement in performance on the post-test relative to the pre-test, referred to as the

‘gain’ score. It can be seen that the gains were in general smallest in the CA-group, due tothe ceiling effect, and were of comparable magnitude in the dyslexic and SA groups.

Both groups improved more in spelling derivations than in spelling base words. The

SA-group achieved approximately equal gains in spelling Change and No Change

derivations while the dyslexic group improved more on the No Change items.

Main training effectsThe first hypothesis stated that gains by CA-group would be smaller than the other two

groups due to their high performance before training. The gains of the dyslexic group

were expected to be comparable to those of younger children of the same spelling level.

The significance of the gains was tested in analyses of variance in which Test (Pre- vs.

Post-) was a within-participants factor and group (Dyslexics, CA-group, SA-group) a

between-participants factor. These verified significant effects for Testing

(F(1,43) ¼ 55.09; p , .001 for the Adjective Study; F(1,42) ¼ 95.68; p , .001 for the

Noun Study) and Group (F(2,43) ¼ 28.27; p , .001 for the Adjective Study;F(2,42) ¼ 28.08; p , .001 for the Noun Study). The interaction testing by group

(F(2,42) ¼ 15.19; p , .001) was significant only for the Noun Study, a difference

attributed to the smaller gains made by the CA-group. Post hoc analysis (Tukey HSD)

revealed that the dyslexic group performed significantly below the CA-group ( p , .001)

and the SA-group ( p , .001 for the Adjectives; p , .01 for the Nouns). The two control

groups also differed ( p , .001). Replication of the analysis on the Dyslexic and SA-

groups confirmed that the small difference in gains was not significant (testing by group

(F(1,30) , 1 for adjectives; (F(1, 29) , 1 for Nouns) indicating that the dyslexicstudents presented the same degree of change following the intervention as the younger

group of the same reading ability. Analysis of variance by Item revealed significant

effects for testing only for the adjectives (F(1,386) ¼ 34.56; p , .001). Group effects

were significant in both studies (F(2,772) ¼ 1015.98; p , .001 for adjectives;

F(2,796) ¼ 992.41; p , .001 for nouns) while the interaction group by testing

(F(2,772) ¼ 13.08; p , .001) was significant only for the Adjective Study.

The size of the potential for gains in spelling ability might be expected to vary

depending on the orthographic level which had been achieved. In order to examine thispoint, we considered the relationship between gain scores and reading and spelling age

in the control groups. The correlation between reading (RA) and spelling age (SA) was

strong and significant in both dyslexic (r ¼ .77, p , .01) and control groups (r ¼ .84,

p , .001). Figures 1a,b and 1c,d plot individual gain scores against reading and spelling

ages respectively for the adjective and noun studies. It can be seen that the relationship

was negative for both reading (r ¼ 2.43, p , .01, r ¼ 2.64, p , .001) and spelling

(r ¼ 2.54, p , .01, r ¼ 2.59, p , .001) for the normally developing readers, indicating

a trend for gains to be numerically larger at the lower levels of reading and spelling skill.This relationship was echoed in the results for the dyslexic individuals (see Figures 1a,b

and 1c,d) whose gain scores fell almost within the ^95% boundaries for the control

groups. This analysis confirms that the spelling of the dyslexic individuals advanced in

response to morphographic instruction in line with orthographic level.

12 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

Training effects on word-typesPre-test assessments verified that spelling of derived words was inferior to spelling

of base words for both studies (F(1,43) ¼ 185.17, p , .001 for adjectives;

F(1,42) ¼ 62.48, p , .001 for nouns). These differences were verified by item analyses(F(3,384) ¼ 22.71; p , .001 for adjectives; F(3,396) ¼ 4.71; p , .01 for nouns).

According to the second hypothesis, the training would increase performance on

derived words in every group. The significance of the gains was tested in analyses of

variance in which test (pre- vs. post-) was a within-participants factor. These confirmed

the effects of base/derived words in both studies (F(1,43) ¼ 185.17; p , .001 for the

Adjective Study; F(1,42) ¼ 62.48; p , .001 for the Noun Study). The interactions of test

by base/derived words were also significant (F(1,43) ¼ 56.64; p , .001 for the Adjective

Study; F(1,42) ¼ 28.06; p , .001 for the Noun Study), supporting the conclusion thatthe gains following the intervention were larger for derivations than for the base words.

A subsidiary analysis indicated that the pattern of gains was very similar in the dyslexic

and SA control groups (group by test by base/derived interaction: F(1,30) ¼ 0.20, ns, for

adjectives; F(1,29) ¼ 4.00, ns, for nouns). Analysis by items confirmed the effects of

base/derived words in both studies (F(3,384) ¼ 22.71; p , .001 for adjectives;

F(3,396) ¼ 4.71; p , .01 for nouns).

Preliminary analyses indicated that spelling was more accurate for No Change items

than for Orthographic Change items (F(1,43) ¼ 241.13; p , .001 for Adjectives;

Spelling adjectives

Spelling age (SA)161514131211109876

Gai

ns (

%)

50

40

30

20

10

0

–10

–20

ControlsDyslexics

(c)

Spelling adjectives (total accuracy)

Reading age (RA)6 10 11 12 13 14 15 16

Gai

ns (

%)

–20–10

01020304050

(a)

Ivor

RonVictor

Norman

Jason

Bernard

DanielWilliam

Gordon

7 8 9

Controls

Dyslexics

ControlsDyslexics

Controls

Dyslexics

(b) Spelling nouns (total accuracy)

Gai

ns (

%)

–20–10

01020304050

IvorRon

WilliamVictor

Gordon

DanielBernardNorman

Jason

Reading age (RA)6 10 11 12 13 14 15 167 8 9

Spelling nouns

Gai

ns (

%)

50

40

30

20

10

0

–10

–20

(d)

Spelling age (SA)6 10 11 12 13 14 15 167 8 9

Figure 1. (a) Gains of the adjective study in relation to reading age; (b) gains of the noun study in

relation to reading age; (c) gains of the adjective study in relation to spelling age; (d) gains of the Noun

Study in relation to spelling age. Scatterplot of dyslexic and control accuracy scores (mean %). The

regression line is based upon the control group data, the outer lines mark the boundaries beyond which

scores are significantly outlying from the control group mean at p , .05.

Morphological spelling training and dyslexia 13

BJP 470—7/10/2008—ANISH—309616

F(1,42) ¼ 144.65; p , .001 for Nouns). These differences were also significant in

item analyses (F(1,190) ¼ 16.43; p , .001 for Adjectives; F(1,196) ¼ 26.10; p , .001

for Nouns).

The third assumption stated that training would enhance performance on

Orthographic Change items for each group. The relative gains for Change and No

Change items were tested initially in a 3 (Dyslexics, CA-group, SA-group) by 2 (pre-,post-) by 2 (Change, No Change) analysis of variance. The three-way interactions were

significant (F(2, 43) ¼ 7.07; p , .01 for adjectives; F(2, 42) ¼ 4.35; p , 0.5 for nouns)

showing different effects of training for each group. A closer examination of Tables 3

and 4 shows that only the CA-group improved their spelling of Change words after

intervention. Another 2 (dyslexic, SA) by 2 (pre-, post-) by 2 (Change, No Change)

analysis of variance showed that the three-way interactions were significant

(F(1,30) ¼ 4.51; p , .05 for adjectives; F(1, 29) ¼ 7.49; p , 0.5 for nouns), supporting

the conclusion that gains on Change items were weaker than for No Change items in thedyslexic group but more or less equivalent in the SA group.

Generalisation effectsThe generalisation of learning was tested by comparing results for the Trained,Untrained 1, and Untrained 2 item sets. If the pre- versus post-test differences

reflect merely a general improvement occurring between the two test points, we

would expect all three sets of items to show similar gains. If, alternatively, the gains

were specific to the items that were directly taught in the intervention, then we

would expect gains to occur for the Trained items but not for either of the

Untrained sets. If the training generalizes across the wider set of morphologically

complex words containing particular suffixes, then we expect the gains to extend

to Untrained 1 derivations but not to Untrained 2 items. Figure 2a and 2b show thepre- and post test accuracy scores for the three training sets for each group in the

two studies. Both data sets favour an account in which there is generalisation of

training from taught items to untaught items containing the same suffixes. The pre-

versus post test gains were large for Trained items, intermediate for Untrained 1

items, and small for Untrained 2 items.

These conclusions were tested in analyses of variance in which groups (dyslexics,

CA-group, SA-group), test (pre-, post-), and training (Trained, Untrained 1, Untrained 2)

were factors. There were significant effects of training set (F(2, 86) ¼ 14.68; p , .001for adjectives; F(2, 84) ¼ 98.45; p , .001 for nouns) and a training by test interaction

(F(2, 86) ¼ 20.21; p , .001 for adjectives; F(2, 84) ¼ 43.54; p , .001 for nouns). This

interaction indicates a significant variation in the size of the pre- versus post test

difference between the three word sets. Further analysis of Trained and Untrained 1

words showed significant training by test interactions (F(1,43) ¼ 18.53; p , .001 for

adjectives; F(1, 42) ¼ 23.15; p , .001 for nouns), confirming that each group gained

more on Trained than on Untrained 1 words. The interaction training by test by group

was not significant (F(2, 43) , 1). Comparison of Untrained 1 and Untrained 2 wordssimilarly gave a significant training by test interaction (F(1,43) ¼ 8.87; p , .01 for

adjectives; F(1, 42) ¼ 31.32; p , .001 for nouns), indicating that each group performed

better on Untrained 1 than on Untrained 2 words. This demonstrates that the groups

generalized effectively from the Trained words to the analogous Untrained 1 words. The

patterns of training effects were very similar in the dyslexic and SA-control groups

14 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

(group by training by test interaction, F(2,60) , 1 for the adjective study, and

F(2,42) ¼ 1.16; ns for the Noun Study).

Long-term effects for the dyslexic groupDurability of training effects was explored by carrying out a delayed post test, after two

months of the completion of the study, only with the dyslexic group. Comparisons

between the pre-test and the delayed post test, and between the post test and delayed

post test (see Tables 3 and 4) established that accuracy on the delayed post test was

significantly higher than on the pre-test (F(1,8) ¼ 25.67; p , .01 for the Adjective Study,F(1,8) ¼ 12.51; p , .01 for the Noun Study). This effect was significant in an item

analysis of the adjectives data (F(1,387) ¼ 23.18; p , .001) but not in the nouns data

(F(1,399) , 1). The reduction in accuracy between the post test and the delayed post

test in the Adjective Study was not significant (F(1,8) , 1). However, there was a loss of

Training on derived words (adjective study)

Acc

urac

y (%

)

0

20

40

60

80

100(a)

DyslexicsCA-groupRA-group

DyslexicsCA-groupRA-group

Pre Post Pre Post Pre PostTrained Untrained 1 Untrained 2

Pre Post Pre Post Pre PostTrained Untrained 1 Untrained 2

Training on derived words (noun study)

Acc

urac

y (%

)

0

20

40

60

80

100(b)

Figure 2. Training of derived words: (a) The Adjective Study; (b) The Noun Study.

Morphological spelling training and dyslexia 15

BJP 470—7/10/2008—ANISH—309616

about five percentage points in the Noun Study (F(1,8) ¼ 8.06; p , .05). A separate

analysis of the results for spelling derivations suggested significant retention in the

Adjective Study (pre- vs. delayed post test difference, F(1,8) ¼ 21.92; p , .01 by

participants, F(1,193) ¼ 52.37; p , .001 by items) but not in the Noun Study

(F(1,8) ¼ 1.03 by participants, F(1,199) , 1 by items).

Quality of errorsPrevious studies suggest that the spelling of a derivation depends on the accuracy of

spelling the base word (Arnbak & Elbro, 1998; Kemp, 2006). This issue was explored byQ1

‘post hoc’ analyses classifying the word pairs according to whether the base word or the

derivation was misspelled (B þ D- type: e.g. mist-mistie, and B 2 D- type: e.g. suczes-

suzessfol for success-successful ) and examining the location of errors in attempts atspelling derivations (the stem, the suffix, the orthographic change, or a combination of

loci). The outcomes are summarized in Table 5.

The dyslexic students differed from the controls ( p , .001) by having the majority of

their misspelled word pairs in the B 2 D- type. In the pre-tests, the preponderance of

B þ D- errors affected either the suffix ( p , .001) or the critical letters which changed

due to derivational rules (Rule errors). Training effects, especially for the dyslexic group,

involved a reduction in errors on the suffix of No Change derivations ( p , .001). This

was also true of the Orthographic Change items, with the difference that the dyslexicgroup decreased their errors on the suffix (10.50%) but presented moderate losses on

the rule (2.70%), whereas the other groups made higher gains on the rule (about 8%)

and lesser gains on the suffix (about 4%). Closer inspection showed that dyslexic

students particularly confused the ‘magic –e’ with the ‘change -y- to -i-’ rule (see also

Carlisle, 1987). Given that these two rules apply to the majority of Orthographic Change

items, this may explain the dyslexics’ failure to make progress on these items. This

clarifies the contrasting pattern between the dyslexic and control groups which

emerged in relation to the morphological complexity factor. The dyslexic grouppresented an increase in accuracy on the No Change items, while the other two groups

presented more gains on Orthographic Change items.

The pre-test data for B 2 D- pairs showed that errors were located primarily on the

stem or on additional parts of the word. A contrasting pattern between the dyslexic

group and the older group again emerged. The dyslexic students, like the younger group

(especially the children with the lowest spelling ages), presented errors scattered across

the entire word (Across Word errors) (42% for dyslexic group, 23% for SA-group) and at a

smaller percentage on the stem (15 and 18%, respectively). By contrast, the older grouppresented errors mainly on the stem of the derived words (10%) (see Table 5). This

reinforces the view that the dyslexic students were unable to produce the spellings of a

derived word because they lacked essential information about the base form. Their

difficulty in spelling the base word was added to their insufficient knowledge of the

suffix, resulting in derived words which were particularly distorted in terms of their

internal morphological organisation. Examination of the phonological regularity of

these errors (Tsesmeli, 2002) by the dyslexic and the younger group showed quite

severe deviations from their target sound (about 20 and 10%, respectively for responsessuch as corerent-cinerg for co-operate/cooperation). Intervention had a profound effect

on these errors. Improved spelling of the suffix (9.50% for dyslexics) of the words was

accompanied by reduced errors on the stem or across the word (20 and 26% for

dyslexics). There was also a reduction in severe deviations from the target (from 20 to

16 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

Table

5.Pre-testerrorrate

oftheadjectiveandnounstudies

Adjectives

Nouns

Stem

(%)

Suffix(%)

Word

(%)

Stem

(%)

Suffix(%)

Word

(%)

NC(B

þD-)

Dyslexics

1.48(1.54)

21.48(12.11)

3.33(2.63)

3.14(3.37)

8.51(6.94)

2.59(3.23)

CA-group

1.77(1.83)

6.22(5.25)

0.11(0.43)

3.09(3.38)

1.78(1.90)

0.11(0.44)

SA-group

1.88(1.90)

13.18(8.61)

1.95(3.46)

2.95(2.71)

8.63(7.77)

1.96(2.80)

Dyslexics

18.14(11.85)

1.11(1.44)

21.29(16.47)

18.70(7.06)

2.22(2.50)

26.29(17.88)

CA-group

4.22(3.26)

0.11(0.43)

0.88(1.76)

2.85(2.72)

0.11(0.44)

0.11(0.44)

SA-group

17.60(9.90)

0.36(0.99)

11.73(14.82)

12.42(7.12)

0.22(0.58)

12.50(12.29)

Adjectives

Nouns

OC(B

þD-)

Stem

(%)

Rule(%)

Suffix(%)

Word

(%)

Stem

(%)

Rule(%)

Suffix(%)

Word

(%)

Dyslexics

0.00(0.00)

4.34(5.70)

12.24(6.56)

2.48(2.55)

1.48(1.54)

6.48(5.09)

6.11(4.85)

0.74(0.87)

CA-group

1.85(2.66)

5.67(6.11)

6.81(7.57)

0.55(1.49)

2.97(4.98)

4.64(6.64)

1.19(1.65)

0.00(0.00)

SA-group

1.15(1.67)

8.72(7.23)

9.70(6.63)

1.00(1.38)

1.51(1.84)

7.27(6.20)

7.87(5.61)

0.90(1.51)

OC(B-D-)

Dyslexics

15.07(9.61)

0.83(0.99)

8.20(5.73)

42.61(25.76)

22.22(7.07)

1.66(1.66)

4.44(3.43)

31.11(18.37)

CA-group

10.52(4.16)

0.11(0.43)

0.11(0.43)

2.30(3.16)

7.26(5.49)

0.11(0.44)

0.11(0.44)

0.11(0.44)

SA-group

18.89(11.48)

0.76(1.34)

3.41(3.83)

23.29(24.37)

21.81(11.50)

0.75(1.23)

2.95(3.16)

13.33(15.88)

Groups’means(inper

cent)in

term

sofMorphologicalComplexity(N

C,OC)andErrorPair-type(B

þD-,B-D

-)(standarddeviationsin

parentheses).

Morphological spelling training and dyslexia 17

BJP 470—7/10/2008—ANISH—309616

2% for dyslexics), resulting in misspellings of improved quality (e.g. corerent-cinerg

resulted to coloprat-coloprative). Arnbak and Elbro (2000) also reported that trained

dyslexics gained more than untrained ones on phonetically acceptable misspellings

even though the two groups did not differ in terms of phonological skills. A reduction in

severe deviations may be considered a significant consequence of the intervention,

given that the phonetic re-organisation occurred on Trained and Untrained butanalogous (Untrained 1) items ( p , .001) but not on untrained and unrelated

(Untrained 2) items. In any case, the intervention benefitted suffix spelling (also in

Arnbak & Elbro, 2000) and this produced indirect effects on the overall orthographic

form of the word. For instance, when a student knew how to spell the base word

appropriately, gains on the suffix led to a completely accurate spelling (luck-luckie to

luck-lucky). In cases where the student did not know the base word, there was a gain in

the quality of the response (veery-veeryes to very-verious for the target vary-various).

General discussion

This exploratory study aimed to determine how far explicit training of

morphological structure could improve the spelling of derived words by dyslexic

adolescents. In general, the outcomes encourage the view that morphographic

training of this type is a potentially useful approach to the treatment of severe

spelling difficulties.At the outset, the members of the dyslexic group were severely impaired in

spelling both base words and derivations and produced poorly structured responses

containing multiple errors. The intervention improved spelling accuracy, especially

for complex derivations. Gains were relatively small in the CA group, due to the

ceiling effect, but substantial and of comparable magnitude in the dyslexic and SA

control groups. This outcome reinforces the argument that the gains in spelling

came from the training of morphological structure and not simply from spelling

practise. Students received the same amount of practise on base and derived wordsbut improved more dramatically on the derived words which were explicitly

targeted in the intervention. These results are consistent with other findings by

Arnbak and Elbro (2000), and Nunes et al. (2003) who found that morphological

training effects were stronger on the spelling of morphologically complex words

(i.e. derivations, compounds). In terms of morphological complexity factor, pre-test

data showed that each group was worse on words involving orthographic changes

(cf. Leong, 1989, 2000; Leong & Parkinson, 1995; Moats, 2000), but only the CA-Q1

group improved spelling of these words after intervention. The dyslexic groupimproved more on the No Change items, while the SA-group achieved

approximately equal gains in spelling Change and No Change derivations. Learning

was not item-specific but generalized to structurally analogous words which had not

been taught, and this effect was present for each group (cf. Freyd & Baron, 1983;Q1

Wysocki & Jenkins, 1987). There was evidence of long-term retention for the

dyslexic group from the delayed post test, especially in the Adjective Study.

These results have implications for theoretical accounts of the way in which

competence in spelling develops, especially the suggestion that phonographic andlexigraphic foundations precede the level of morphography. If this cumulative account

was valid, we would expect to find that morphographic training was less effective at

lower than at higher reading and spelling ages. This is not what was found,

since gain scores by the control participants were strongest at the lower age levels

18 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

(see Figure 1a–d) and in the most severely impaired dyslexic individuals. Despite a lack

of phonographic and lexigraphic foundations, as shown by the high error rates on both

base and derived words and by responses containing multiple phonetic deviations, the

dyslexic adolescents improved their spelling following the intervention. The outcome

suggests a retroactive effect in which instruction targeted at the morphographic level

produced improvements at the lower levels indexed by reductions in severely deviantresponses and multiple errors.

The most plausible explanation may be stated in terms of a process of

unitisation. Learning to spell long and complex words is facilitated if it is possible to

segment the phoneme–grapheme array into manageable and coherent units. These

units might be defined in terms of phonology, syllables for example, or, as in the

present study, in terms of morphology, as word stems and affixes. For English, the

syllable may not be such a useful structure because syllable boundaries are often

ambiguous and because of varying stress assignment. Further, the fundamentaldifficulty in dyslexia is held to focus on phonological segmentation (Snowling,

2000). This opens the possibility that the morpheme may offer a viable alternative

principle of unitisation for dyslexic individuals. For this to work, it seems necessary

that morphological segmentation should be relatively unimpaired in dyslexia.

Whether this is so is unclear since several studies have indicated that performance

on morphological awareness tasks is deficient relative to chronological age controls

(Bryant, Nunes, & Bindman, 1998; Carlisle, 1987; Elbro, 1990; Fowler & Liberman,

1995; Leong, 1989; Leong & Parkinson, 1995). However, Tsesmeli and Seymour(2006) recently reported an additional study with the present sample in which

morphological awareness was assessed. The results for the dyslexic group were well

below those of the CA-control but exactly in line with reading and spelling age,

implying that morphological awareness might be a product of orthographic

development. This is consistent with a theory in which experience with written

language leads to the development of an ‘orthographic awareness’ of positionally

constrained and recurrent letter sequences. In the normal course, these come to be

perceived as corresponding to free and bound morphemes and this results in theemergence of explicit morphological awareness. This does not happen in individuals

with dyslexia because the instability and imprecision of the spelling system prevents

the isolation of recurring orthographic sequences. The effect of the intervention

may be to provide an alternative route to morphographic awareness through the

use of colour coding and other techniques.

Morphographic training provides a basis for segmentation of long and difficult words

into units of manageable size, therefore, and this assists the understanding of the

phonological basis of spelling and the learning of word- and morpheme-specificspellings. The additional aspect of rule-based alterations required for correct spelling of

some derivations appeared to be more difficult to establish. Young normal spellers

gained from instruction in these rules but this was less true of the dyslexic spellers who

continued to find the application of the rules confusing and difficult. Possibly this rule-

based aspect is the key feature of the morphographic level of spelling and one which

remains somewhat beyond the reach of dyslexic individuals.

Although the main scope of this study is focused on the derivational morphology

in a deep orthography such as English, there is the assumption, despite the lackof appropriate cross-linguistic studies, that these processes might follow the same

course in other more transparent languages (i.e. German or Greek). Since it is now

recognized that in consistent orthographies phonological coding poses less difficulty

Morphological spelling training and dyslexia 19

BJP 470—7/10/2008—ANISH—309616

(Wimmer, Landerl, & Frith, 1999), the acquisition of morphography may be acquired

earlier in the developmental trajectory. For instance, Tsesmeli (2007) found that

Greek data supported an earlier acquisition of morphology in comparison with

English. Spelling of derived words was statistically lower than their bases also in

Greek, but the discrepancy between base and derived words were smaller than in

English (13% vs. 22%). Qualitative analyses showed that stem consistency indexes bystudents of lower reading/spelling ages were significantly higher in Greek than in

English (70% vs. 50%). More importantly, English children presented a higher

percentage of serious errors (47%) which were obscuring the morphological

structure of words in comparison with Greek children (18%), for whom the majority

of errors appeared either on the stem (27%) or on the suffix (17%) thus retaining the

inner word structure intact.

Some of the limitations of this study would be that all stages of this intervention were

implemented by the same person and this may affect the validity of the results due to a‘teacher effect’. However, instruction led to a variation to the spelling performance of

each group according to their reading and spelling ability across conditions, and these

results are consistent to experimental literature. The specificity of the morphological

training effects also could be further verified by extraneous control groups, for example,

groups which were not trained only to morphology (e.g. phonology or reading

comprehension) or even untrained dyslexic groups matched in terms of chronological

age and cognitive measures as in other intervention studies (Bryant, Nunes, & Bindman,

2000; Elbro & Arnbak, 1996). In addition, in future experimental designs of similarinterventions would have to be controlled more tightly for factors that would affect

the results (i.e. word length or the phonological complexity of items) or to examine the

co-occurrence of phonological and morphological factors as different aspects of the

same words integrated in one experimental design, as suggested by Carlisle (1995).

However, very few studies (Jones, 1991; Rubin, 1988) offered systematic comparisons

of the awareness of phonological and morphological structure of different types of

words and more studies of this kind would be particularly important for the complex

nature of this area. More interestingly, an investigation of morphophonology, referringto the discrete changes occurring between the morpheme boundaries within the word,

would be particularly revealing for the nature of the deficiencies of the dyslexic

students.

Nonetheless, the study suggests that the training of morphological structure can

be of general benefit to the spelling of dyslexic adolescents (see also Bryant et al.,

1998; Elbro & Arnbak, 1996). The goal of such training is to provide the student

with the skills necessary to carry out morphological decomposition of complex

words. The important feature of the present procedure was the adoption of amorphographic approach to decomposition which aimed to encourage awareness

of the division of written words into subsets of letters corresponding to free and

bound morphemes.

Acknowledgements

We are most grateful to Perth & Kinross City Council and to the students and staff of Perth High

School and Caledonian Primary School in Scotland, UK for their generous assistance and time. The

research was part of a doctoral dissertation conducted by S. N. Tsesmeli and supported financially

by the Greek State Scholarships Foundation.

20 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

References

Arnbak, E., & Elbro, C. (2000). The effects of morphological awareness training on the reading and

spelling skills of young dyslexics. Scandinavian Journal of Educational Research, 44(3),

229–251.

Backman, J. E., Mamen, M., & Ferguson, H. B. (1984). Reading level design: Conceptual and

methodological issues in reading research. Psychological Bulletin, 96, 560–568.

British Psychological Society (1999). Dyslexia, literacy and psychological assessment. Leicester:

The BPS.

Bruck, M. (1992). Persistence of dyslexics’ phonological awareness deficits. Developmental

Psychology, 28, 874–886.

Bruck, M. (1993). Component spelling skills of college students with a childhood diagnosis of

dyslexia. Learning Disabilities Quarterly, 16, 171–184.

Bryant, P. E., & Bradley, L. (1985). Children’s reading problems. Oxford: Blackwell.

Bryant, P., Nunes, T., & Bindman, M. (1998). Awareness of language in children who have reading

difficulties. Historical comparisons in a longitudinal study. Journal of Child Psychology and

Psychiatry, 39(4), 501–510.

Bryant, P., Nunes, T., & Aidinis, A. (1999). Different morphemes, same spelling problems:

Cross-linguistic developmental studies. In M. Harris & G. Hatano (Eds.), Learning to read

and write: A cross-linguistic perspective (pp. 112–133). Cambridge: Cambridge University

Press.

Bryant, P., Nunes, T., & Bindman, M. (1999). Morphemes and spelling. In T. Nunes (Ed.), Learning

to read: An integrated view from research and practise (pp. 15–41). Dordrecht, Netherlands:

Kluwer Academic Publishers.

Bryant, P., Nunes, T., & Bindman, M. (2000). The relations between children’s linguistic awareness

and spelling: The case of the apostrophe. Reading and Writing: An Interdisciplinary Journal,

12, 253–276.

Byrne, B., & Fielding-Barnsley, R. (1995). Evaluation of a program to teach phonemic awareness to

young children. Journal of Educational Psychology, 83, 451–455.

Carlisle, J. F. (1987). The use of morphological knowledge in spelling derived forms by learning-

disabled and normal students. Annals of Dyslexia, 37, 90–108.

Carlisle, J. F. (1995). Morphological awareness and early reading achievement. In L. B. Feldman

(Ed.), Morphological aspects of language processing. New Jersey: Lawrence Erlbaum

Associates.

Carrol, J. B., Davies, P., & Richman, B. (1971). Word frequency book. New York, NY: American

Heritage Publishing Co.

Critchley, M. (1981). Dyslexia: An overview. In G. Th. Pavlidis & T. R. Miles (Eds.), Dyslexia

research and its applications to education (pp. 1–11). Chichester: Wiley & Sons.

Deacon, S. H., Parrila, R., & Kirby, J. R. (2006). Processing of derived forms in high functioning

dyslexics. Annals of Dyslexia, 56(1), 103–128.

Derwing, B. L. (1976). Morpheme recognition and the learning of rules for derivational

morphology. The Canadian Journal of Linguistics, 21(1), 38–66.

Derwing, B. L., Smith, M. L., & Wiebe, G. E. (1995). On the role of spelling in morpheme

recognition: Experimental studies with children and adults. In L. B. Feldman (Ed.),

Morphological aspects of language processing (pp. 3–27). Mahwah, NJ: Erlbaum.

Department for Education (1994). Code of practise on the identification and assessment of

special educational needs. London: HMSO.

Elbro, C. (1990). Differences in dyslexia: A study of reading strategies and deficits in a linguistic

perspective. Munksgaard: Copenhagen.

Elbro, C., & Arnbak, E (1996). The role of morpheme recognition and morphological awareness in

dyslexia. Annals of Dyslexia, 46, 209–240.

Elliott, C. D. (1992). British Ability Scales. Spelling scale. Manual. Windsor, UK: NFER-Nelson.

Elliott, C. D., Murray, D. J., & Pearson, L. S. (1983). British Ability Scales. Windsor: NFER-Nelson

Publishing Company Ltd.

Morphological spelling training and dyslexia 21

BJP 470—7/10/2008—ANISH—309616

Fowler, A. E., & Liberman, I. Y. (1995). The role of phonology and orthography in morphological

awareness. In L. B. Feldman (Ed.), Morphological aspects of language processing

(pp. 157–188). Mahwah, NJ: Erlbaum.

Frith, U. (1985). Beneath the surface of developmental dyslexia. In K. Patterson, M. Colteheart, &

J. Marshall (Eds.), Surface dyslexia. London: Lawrence Erlabaum.

Freyd, P., & Baron, J. (1982). Individual differences in acquisition of derivational morphology.

Journal of Verbal Learning and Verbal Behavior, 21, 282–295.

Hatcher, P. J., Hulme, C., & Ellis, A. W. (1994). Ameliorating early reading failure by integrating the

teaching of reading and phonological skills: The phonological linkage hypothesis. Child

Development, 65, 41–57.

Henry, M. K. (1988). Beyond phonics: Integrated decoding and spelling instruction based on word

origin and structure. Annals of Dyslexia, 38, 258–275.

Henry, M. K. (1990). Words: Integrated decoding and spelling instruction based on word origin

and structure. Los Gatos, CA: Lex Press.

Henry, M. K. (1993). Morphological structure: Latin and Greek roots and affixes as upper grade

code strategies. Reading and Writing: An Interdisciplinary Journal, 5, 227–241.

Hoien, T., & Lundberg, I. (2000). Dyslexia. From theory to intervention. Dordrecht, NL: Kluwer

Academic Publishers.

Jones, N. K. (1991). Development of morphophonemic segments in children’s mental

representations of words. Applied Psycholinguistics, 12, 217–239.

Katamba, F. (1993). Morphology. London: MacMillan Press Ltd.

Kemp, N. (2006). Children’s spelling of base, inflected, and derived words: Links with

morphological awareness. Reading and Writing: An Interdisciplinary Journal, 19,

737–765.

Lefly, D. L., & Pennington, B. F. (1991). Spelling errors and reading fluency in compensated adult

dyslexics. Annals of Dyslexia, 41(1), 141–162.

Leong, C. K. (1989). Productive knowledge of derivational rules in poor readers. Annals of

Dyslexia, 39, 94–115.

Leong, C. K. (2000). Rapid processing of base and derived forms of words in grades 4, 5 and 6

children’s spelling. Reading and Writing: An Interdisciplinary Journal, 12, 277–302.

Leong, C. K., & Parkinson, M. E. (1995). Processing of English morphological structure by

poor readers. In C. K. Leong & R. M. Joshi (Eds.), Developmental and acquired

dyslexia. Neuropsychological and neurolinguistic perspectives (pp. 237–261). Dordrecht,

Netherlands: Kluwer Academic Publishers.

Liberman, I. Y., Shankweiler, D., & Liberman, A. (1989). The alphabetic principle and learning to

read. In D. Shankweiler & I. Y. Liberman (Eds.), Phonology and reading disability. Ann Arbor:

The University of Michigan Press.

Lundberg, I., Frost, J., & Petersen, O. (1988). Effects of an extensive program for stimulating

phonological awareness in pre-school children. Reading Research Quarterly, 23, 263–284.

Mamen, M., Ferguson, H. B., & Backman, J. E. (1986). No difference represents a significant

finding: The logic of the reading level design. A response to Bryant and Goswami.

Psychological Bulletin, 100, 104–106.

Moats, L. C. (1998). Intervention research in dyslexia. Perspectives (The International Dyslexia

Association), 24(4), 22–23.

Nagy, W. E., & Anderson, R. C. (1984). How many words are there in printed school English?

Reading Research Quarterly, 19, 304–330.

Nunes, T., Bryant, P., & Olsson, J. (2003). Learning morphological and phonological spelling rules:

An intervention study. Scientific Studies of Reading, 7(3), 289–307.

Pennington, B. F., McCabe, L. L., Smith, S. D., Lefly, D. L., Bookman, M. O., Kimberling, W. J., et al.

(1986). Spelling errors in adults with a form of familial dyslexia. Child Development, 57(4),

1001–1013.

Pennington, B. F., Van Orden, G. C., Smith, S. D., Green, P. A., & Haith, M. (1990). Phonological

processing skills and deficits in adult dyslexics. Child Development, 61, 1753–1778.

22 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

Raven, J. C. (1958). Standard progressive matrices (Sets A, B, C, D and E). London: H.K. Lewis.

Raven, J. C. (1962). Coloured progressive matrices (Sets A, AB and B). London: H.K. Lewis.

Rubin, H. (1988). Morphological knowledge and early writing ability. Language and Speech,

31(4), 337–355.

Seymour, P. H. K., & Evans, H. M. (1993). The visual (orthographic) processor and developmental

dyslexia. In D. Willows, R. Kruk, & E. Corcos (Eds.), Visual processes in reading and reading

disabilities. Hillsdale, NJ: Lawrence Erlbaum Associates.

Seymour, P. H. K., & Duncan, L. G. (2001). Learning to read in English. Psychology: The Journal of

the Hellenic Psychological Society, 8(3), 281–299.

Senechal, M., Basque, M. T., & Leclaire, T. (2006). Morphological knowledge as revealed in

children’s spelling accuracy and reports of spelling strategies. Journal of Experimental Child

Psychology, 95, 231–254.

Senechal, M., & Kearnan, K. (2007). The role of morphology in reading and spelling. Advances in

Child Development and Behavior, 35, 297–325.

Snowling, M. (2000). Dyslexia (2nd edn). UK: Blackwell Publishers.

Torgesen, J. K., & Wagner, R. K. (1992). Language abilities, reading acquisition and developmental

dyslexia. Journal of Learning Disabilities, 25, 577–581.

Tsesmeli, S. N. (2002). Derivational morphology, spelling ability and dyslexia in secondary school

students: Assessment and intervention. Unpublished doctoral dissertation, University of

Dundee, Scotland.

Tsesmeli, S. N. (2007). The role of derivational morphology in the development of spelling ability:

a cross-linguistic approach (in Greek). Psychology: The Journal of the Hellenic Psychological

Society, 14(3), 231–249.

Tsesmeli, S. N., & Seymour, P. H. K. (2006). Derivational morphology and spelling in dyslexia.

Reading and Writing: An Interdisciplinary Journal, 19(6), 587–625.

Wimmer, H., Landerl, K., & Frith, U. (1999). Learning to read German: Normal and impaired

acquisition. In M. Harris & G. Hatano (Eds.), Learning to read and write. A cross-linguistic

perspective (pp. 34–50). Cambridge, UK: Cambridge University Press.

Wysocki, K., & Jenkins, J. R. (1987). Deriving word meanings through morphological

generalization. Reading Research Quarterly, 22(1), 66–81.

Received 29 January 2007; revised version received 8 September 2008

Appendix I. Dyslexic participants

Dyslexics Sex CA BAS-RA Centiles BAS-SA Centiles Raven’s SPM Digit span

1 M 14.05 7.4 2 7.2 1 25 42 M 13.04 14.0 70 10.10 30 50–75 673 M 13.11 7.9 4 7.9 4 50 334 M 14.04 11.7 41 9.1 12 25–50 495 M 13.08 9.4 14 9.1 12 50–75 26 M 15.03 8.11 8 7.6 2 50–75 277 M 14.09 9.5 12 7.2 1 10–25 88 M 14.11 9.1 14 8.1 4 50–75 109 M 14.11 8.3 5 9.1 9 50 46

Sex, chronological age (CA), BAS-reading age (RA)/spelling age (SA) and centiles, standard progressivematrices (SPM) (percentiles) and digit span (centiles).

Morphological spelling training and dyslexia 23

BJP 470—7/10/2008—ANISH—309616

Appendix

II.TheAdjectiveStudy:Item

sanddesign

Nochange

item

sOrthographicchange

item

s

Bases

T/U

fl

Derives

T/U

fLf

lBases

T/U

fl

Derives

T/U

fLf

l

luck

T193

4Lucky

TH

166

5fun

T695

3funny

TH

312

5rain

T938

4Rainy

TH

122

5sun

T1977

3sunny

TH

116

5dust

T340

4Dusty

TH

79

5noise

T411

5noisy

TH

94

5rust

U39

4Rusty

UH

55

5taste

U283

5tasty

UH

29

5thirst

U33

6Thirsty

UH

55

7fog

U212

3foggy

UH

37

7trust

T103

5Trusty

TL

46

wave

T318

4wavy

TL

26

4mist

T57

4Misty

TL

19

5rose

T461

4rosy

TL

21

4crust

U160

5Crusty

UL

46

fat

U386

3fatty

UL

13

5fault

U120

5Faulty

UL

10

6spot

U403

4spotty

UL

56

mess

U57

4Messy

UL

12

5breeze

U147

6breezy

UL

56

use

T7009

3useful

TH

430

6beauty

T330

6beautiful

TH

1048

9help

T3875

4helpful

TH

209

7plenty

T320

6plentiful

TH

56

9success

T242

7successful

TH

236

10

pity

T56

4pitiful

TH

13

7wonder

U445

6wonderful

UH

411

9fancy

U103

5fanciful

UH

16

8harm

U148

4harmful

UH

94

7–

–thank

T301

5thankful

TL

45

8bounty

T9

6bountiful

TL

69

pain

T198

4painful

TL

36

7duty

T147

4dutiful

TL

27

truth

U215

5truthful

UL

98

mercy

U49

5merciful

UL

28

trust

U103

5trustful

UL

28

––

respect

U185

7respectful

UL

16

10

––

danger

T359

6dangerous

TH

308

9fame

T85

4famous

TH

717

6mountain

T838

8mountainous

TH

87

11

vary

T108

4various

TH

620

7poison

T76

6poisonous

TH

77

9marvel

T23

6marvellous

TH

56

10

prosper

U11

7prosperous

UH

43

10

glory

U93

5glorious

UH

54

8thunder

U140

7thunderous

UH

17

10

grace

U55

5gracious

UH

27

8mom

ent

T834

6mom

entous

TL

10

9adventure

T152

9adventurous

TL

23

11

murder

T40

6murderous

TL

59

luxury

T25

6luxurious

TL

17

9pomp

U13

4pompous

UL

57

space

U1499

5spacious

UL

17

8

24 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

Appendix

II.(Continued)

Nochange

item

sOrthographicchange

item

s

Bases

T/U

fl

Derives

T/U

fLf

lBases

T/U

fl

Derives

T/U

fLf

l

rigor

U2

5rigorous

UL

38

desire

U129

6desirous

UL

08

ruin

U43

4ruinous

UL

07

envy

U24

4envious

UL

97

Act

T457

3active

TH

178

6expense

T38

7expensive

TH

129

9protect

T340

7protective

TH

70

10

create

T235

6creative

TH

58

8effect

T399

6effective

TH

136

9narrate

T2

7narrative

TH

39

9attract

U101

7attractive

UH

128

10

execute

U13

7executive

UH

50

9detect

U32

6detective

UH

67

9decorate

U29

8decorative

UH

25

10

digest

T23

6digestive

TL

40

9intense

T58

7intensive

TL

13

9progress

T228

8progressive

TL

33

11

attribute

T14

9attributive

TL

011

collect

U189

7collective

UL

30

10

imitate

U51

7imitative

UL

09

connect

U92

7connective

UL

30

10

co-operate

U21

9co-operative

UL

23

11

select

U193

6selective

UL

10

9manipulate

U4

10

manipulative

UL

012

nation

U510

6national

UH

327

8nature

U462

6natural

UH

739

7coast

U594

5coastal

UH

114

7arrive

U136

6arrival

UH

62

7profession

U31

10

professional

UH

122

12

navy

U43

4naval

UH

50

5tradition

U66

9traditional

UH

90

11

culture

U146

7cultural

UH

51

8form

U2720

4form

alU

H85

6survive

U87

7survival

UH

45

8region

U825

6regional

UL

24

8propose

U20

7proposal

UL

22

8season

U370

6seasonal

UL

24

8appraise

U3

8appraisal

UL

49

post

U253

4postal

UL

12

6rehearse

U6

8rehearsal

UL

16

9margin

U59

6marginal

UL

10

8remove

U215

6removal

UL

28

7rent

U71

4rental

UL

46

revive

U11

6revival

UL

77

Key:TforTrained

item

s,U

forUntrained

item

s,fforFrequency,lforLetter-length,H

forHighFrequency,LforLo

wFrequency.

Note

1.Item

sin

italicsinvolvephonologicalchanges(tone-shifts);Theitem

snation-national,space-spacious,nature-naturalinvolvevowelalternations.

Note

2.Meanfrequencies

andstandarddeviations:Bases:358.39(869.86),Derivations:90.77(170.17),(C

arrol,Davies,&Richman,1971).

Morphological spelling training and dyslexia 25

BJP 470—7/10/2008—ANISH—309616

Appendix

III.TheNounStudy:Item

sanddesign

Nochange

item

sOrthographicchange

item

s

Bases

T/U

fL

Derives

T/U

fLf

lBases

T/U

fl

Derives

T/U

fLf

I

dark

T19

4Darkness

TH

250

8happy

T774

5happiness

TH

79

9ill

T148

3Illness

TH

78

7em

pty

T384

5em

ptiness

TH

79

sad

T309

3Sadness

TH

36

7Ugly

T126

4ugliness

TH

58

sick

U334

4Sickness

UH

59

8ready

U1207

5readiness

UH

12

9weak

U245

4Weakness

UH

35

8Lazy

U102

4laziness

UH

48

Fit

T461

3Fitness

TL

10

7steady

T243

6steadiness

TL

210

wild

T929

4W

ildness

TL

10

8dizzy

T11

5dizziness

TL

09

bitter

U103

6Bitterness

UL

11

10

fuzzy

U17

5fuzziness

UL

39

soft

U669

4So

ftness

UL

17

8clumsy

U50

6clumsiness

UL

010

mad

U170

3madness

UL

14

7fussy

U10

5fussiness

UL

09

perform

T168

7perform

ance

TH

149

11

guide

T347

5guidance

TH

30

8assist

T33

6assistance

TH

39

10

endure

T39

6endurance

TH

26

9allow

T229

5allowance

TH

64

9ignore

T28

6ignorance

TH

22

9resist

U47

6resistance

UH

134

10

insure

U29

6insurance

UH

79

accept

U159

6acceptance

UH

32

10

apply

U192

5appliance

UH

14

9attend

T105

6attendance

TL

10

10

admit

T74

5admittance

TL

210

avoid

T246

5avoidance

TL

09

rely

T42

4reliance

TL

12

8disturb

U22

7disturbance

UL

10

11

observe

U349

7observance

UL

810

annoy

U13

5annoyance

UL

13

9assure

U23

6assurance

UL

12

9accord

U12

6accordance

UL

12

10

comply

U2

6compliance

UL

310

personal

T255

8personality

TH

97

11

active

T178

6activity

TH

164

8popular

T393

7popularity

TH

37

10

dense

T86

5density

TH

104

7real

T910

4reality

TH

56

7secure

T75

6security

TH

56

8similar

U555

7similarity

UH

27

10

intense

U58

7intensity

UH

68

9minor

U233

5minority

UH

27

8diverse

U16

7diversity

UH

18

9familiar

T406

8familiarity

TL

13

11

sensitive

T86

9sensitivity

TL

14

11

original

T403

8originality

TL

12

11

positive

T204

8positivity

TL

010

complex

U206

7complexity

UL

20

10

negative

U210

8negativity

UL

010

26 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

Appendix

III.(Continued)

Nochange

item

sOrthographicchange

item

s

Bases

T/U

fL

Derives

T/U

fLf

lBases

T/U

fl

Derives

T/U

fLf

I

human

U710

5humanity

UL

22

8pure

U202

4purity

UL

11

6form

alU

85

6form

ality

UL

89

rare

U126

4rarity

UL

46

act

T457

3action

TH

519

6educate

T11

7education

TH

294

9protect

T340

7protection

TH

153

10

relate

T38

6relation

TH

100

8direct

T284

6direction

TH

651

9imitate

T51

7imitation

TH

38

9collect

U189

7collection

UH

197

10

communicate

U87

11

communication

UH

140

13

invent

U73

6invention

UH

151

9create

U235

6creation

UH

49

8adapt

T25

5adaption

TL

27

8promote

T25

7promotion

TL

19

9suggest

T244

7suggestion

TL

40

10

narrate

T2

7narration

TL

10

9digest

U23

6digestion

UL

45

9tolerate

U5

8toleration

UL

310

prevent

U248

7prevention

UL

15

10

illustrate

U119

10

illustration

UL

16

12

correct

U940

7correction

UL

11

10

separate

U416

8separation

UL

18

10

teach

U254

5teacher

UH

836

7drive

U543

5driver

UH

305

6farm

U900

4farm

erU

H414

6write

U9846

5writer

UH

279

6lead

U520

4leader

UH

310

6run

U1473

3runner

UH

50

6read

U3057

4reader

UH

256

6win

U340

3winner

UH

63

6own

U3006

3owner

UH

172

5manage

U88

6manager

UH

85

7bank

U465

4banker

UL

14

6dance

U608

5dancer

UL

31

6clean

U521

5cleaner

UL

32

7sw

imU

2990

4sw

immer

UL

34

7climb

U288

5climber

UL

14

7produce

U586

7producer

UL

30

8train

U556

5trainer

UL

18

7exam

ine

U190

7exam

iner

UL

48

publish

U23

7publisher

UL

17

9travel

U814

6traveller

UL

29

Key:Tfortrained

item

s,U

foruntrained

item

s,fforfrequency,lforletter-length,H

forhighfrequency,Lforlow

frequency.

Note

1.Item

sin

italicsinvolvephonologicalchanges(tone-shifts).

Note

2.Meanfrequencies

andstandarddeviations:Bases:447,47(1096,74),74,45(134,57),(C

arrol,Davies,&Richman,1971).

Morphological spelling training and dyslexia 27

BJP 470—7/10/2008—ANISH—309616

Appendix IV. Instructions for the workshop

1. I’d like to introduce two words. Please, look at them carefully and tell me which

part is common between the two words. While the child finds it. Please, draw a

line to show me clearly the two different parts (use of pencil).

CARDS luck lucky (analysis of the word in 2 parts)

2. Now, I’ll show you another pair of words. Please, look at them carefully and tell

me which part is common. Then, please cut and separate the two parts (use of

scissors).

CARDS rain rainy (analysis of the word in 2 parts)

3. I’ll give you two cards. Please, copy exactly the word of the first card in the blank

space of the second card.

CARDS dust : : : : : : y (focus on the stem of the word)

4. I’ll present you another set of cards. Do you remember which was the last part of

our words in previous pairs? Please, write it down in the blank space of the

second card.

CARDS trust trust : : : : : : (focus on the suffix of the word)

5. Now, I’ll give you my last pair of cards. Please, write down a new word that

comes from the first one on the second blank card.

CARDS mist : : : : : : (creation of a new word)

28 S. N. Tsesmeli and P. H. K. Seymour

BJP 470—7/10/2008—ANISH—309616

Author QueriesJOB NUMBER: 470

JOURNAL: BJP

Q1 References Arnbak & Elbro (1998); Bradley and Bryant (1983); Moats (2000);Freyd & Baron (1983); Matthews (1991); Katamba, 1994, have been cited in

text but not provided in the list. Please supply reference details or delete the

reference citations from the text.

Q2 Table 6 is cited in the text but has not been provided. Please remove the citation

or provide the table.

Morphological spelling training and dyslexia 29

BJP 470—7/10/2008—ANISH—309616