View in EDS HTML Full Text PDF Full Text

Reading and Reading-Related Skills in Adults with Dyslexia from Different Orthographic Systems: A Review and Meta-Analysis

Saved in:
Bibliographic Details
Title: Reading and Reading-Related Skills in Adults with Dyslexia from Different Orthographic Systems: A Review and Meta-Analysis
Language: English
Authors: Reis, Alexandra (ORCID 0000-0001-5598-0999), Araújo, Susana, Morais, Inês Salomé, Faísca, Luís
Source: Annals of Dyslexia. Oct 2020 70(3):339-368.
Availability: Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
Peer Reviewed: Y
Page Count: 30
Publication Date: 2020
Document Type: Journal Articles
Reports - Research
Descriptors: Reading Skills, Adults, Dyslexia, Orthographic Symbols, Reading Difficulties, Reading Comprehension, Reading Rate, Phonological Awareness, Writing Difficulties, Symptoms (Individual Disorders), Severity (of Disability), Cognitive Processes
DOI: 10.1007/s11881-020-00205-x
ISSN: 0736-9387
Abstract: An individual diagnosed with dyslexia in childhood typically remains dyslexic throughout his/her life. However, the cognitive profile of adults with dyslexia has been less explored than that of children. This meta-analytic study is intended to clarify three questions: (1) To what extent, and in what manner, do adults with reading difficulties (dyslexia) differ from typical adult readers in measures of reading and writing competence and related cognitive skills?; (2) To what extent do speed measures pose a greater challenge than accuracy measures in an adult population that has already had years of print exposure?; and (3) To what extent does orthographic transparency modulate the reading profile of adults with dyslexia? A total of 178 studies comparing adults with dyslexia and matched controls were reviewed. The results showed that adults with dyslexia exhibited poor performance on almost all reading and writing tasks expressed by very large effect sizes (range 1.735 [less than or equal to] "d" [less than or equal to] 2.034), except for reading comprehension ("d" = 0.729). Deficits in reading- and writing-related variables are also present but with a lower expression (range 0.591 [less than or equal to] "d" [less than or equal to] 1.295). These difficulties are exacerbated for speed measures, especially for word and pseudoword reading, phonological awareness and orthographic knowledge. Orthographic transparency proved to be a significant moderator of dyslexic deficits in word and pseudoword reading, reading comprehension, spelling and phonological awareness, with the expression of the deficits being weaker on transparent--as opposed to intermediate and opaque--orthographies. Overall, the meta-analysis shows that reading and writing difficulties persist in adulthood and are more pronounced in speed measures. Moreover, symptoms are more severe for reading and writing than they are for measures tapping into the cognitive processes underlying reading skills. Orthographic transparency has a significant effect on the manifestation of dyslexia, with dyslexia symptoms being less marked on transparent orthographies. In addition, phonological awareness seems to be a minor problem in adulthood, especially for transparent orthographies.
Abstractor: As Provided
Entry Date: 2020
Accession Number: EJ1275867
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwHXhlE8kAI7NWlFShz0yIoZAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDC8k7q03Rj64_OamNQIBEICBmzuMDHgkyM9iI7YqsgjKKhtPjOP2OhvWl19mWwvg6Gr5YNZpFNNh8l3zzmx6v0IMFLYnCcJnFMPN6jcvo91ahqmrmAA3PBYtlLqjfYHlTdzXH7w_nSzpl7SLXG0nD48drFyGuPMeMr8JsAcuS3C9Nl_bMuPPx9qkCnNFskfh33qfOO1L-LKAPPMA_Au0MrFh5ivmlSmREkItbzps
Text:
  Availability: 1
  Value: <anid>AN0146733574;bm501oct.20;2020Nov02.02:54;v2.2.500</anid> <title id="AN0146733574-1">Reading and reading-related skills in adults with dyslexia from different orthographic systems: a review and meta-analysis </title> <p>An individual diagnosed with dyslexia in childhood typically remains dyslexic throughout his/her life. However, the cognitive profile of adults with dyslexia has been less explored than that of children. This meta-analytic study is intended to clarify three questions: (<reflink idref="bib1" id="ref1">1</reflink>) To what extent, and in what manner, do adults with reading difficulties (dyslexia) differ from typical adult readers in measures of reading and writing competence and related cognitive skills?; (<reflink idref="bib2" id="ref2">2</reflink>) To what extent do speed measures pose a greater challenge than accuracy measures in an adult population that has already had years of print exposure?; and (<reflink idref="bib3" id="ref3">3</reflink>) To what extent does orthographic transparency modulate the reading profile of adults with dyslexia? A total of 178 studies comparing adults with dyslexia and matched controls were reviewed. The results showed that adults with dyslexia exhibited poor performance on almost all reading and writing tasks expressed by very large effect sizes (range 1.735 ≤ d ≤ 2.034), except for reading comprehension (d = 0.729). Deficits in reading- and writing-related variables are also present but with a lower expression (range 0.591 ≤ d ≤ 1.295). These difficulties are exacerbated for speed measures, especially for word and pseudoword reading, phonological awareness and orthographic knowledge. Orthographic transparency proved to be a significant moderator of dyslexic deficits in word and pseudoword reading, reading comprehension, spelling and phonological awareness, with the expression of the deficits being weaker on transparent—as opposed to intermediate and opaque—orthographies. Overall, the meta-analysis shows that reading and writing difficulties persist in adulthood and are more pronounced in speed measures. Moreover, symptoms are more severe for reading and writing than they are for measures tapping into the cognitive processes underlying reading skills. Orthographic transparency has a significant effect on the manifestation of dyslexia, with dyslexia symptoms being less marked on transparent orthographies. In addition, phonological awareness seems to be a minor problem in adulthood, especially for transparent orthographies.</p> <p>Keywords: Dyslexic adults; Meta-analysis; Orthographic transparency; Phonological awareness; Reading; Spelling</p> <p>Electronic supplementary material The online version of this article (10.1007/s11881-020-00205-x) contains supplementary material, which is available to authorized users.</p> <hd id="AN0146733574-2">Introduction</hd> <p>An individual diagnosed with dyslexia in childhood typically remains dyslexic throughout his/her life (Hatcher, Snowling, & Griffiths, [<reflink idref="bib24" id="ref4">24</reflink>]; Pammer, [<reflink idref="bib39" id="ref5">39</reflink>]). Affected individuals face difficulties acquiring reading and related cognitive skills—difficulties that persist into adulthood. Therefore, such individuals are at risk of developing secondary emotional and behavioural problems associated with educational failure and later may encounter unemployment and consequent psychological, economic and social problems (Gerber, [<reflink idref="bib21" id="ref6">21</reflink>]; Watson & Boman, [<reflink idref="bib65" id="ref7">65</reflink>]).</p> <p>Despite this life-long persistence, the manifestations of dyslexia in adults, as opposed to children, are far from understood, and the study of adults with dyslexia might bring new insights into the field of developmental dyslexia. While studies with children are critical to understanding the development of early reading skills and the cognitive capacities that predict reading, they do not provide information about the long-term stability of reading deficits and the actual profile of adults with dyslexia. For example, Miller-Shaul's ([<reflink idref="bib35" id="ref8">35</reflink>]) results suggested that, in an opaque orthography, some deficits present in children with dyslexia are attenuated in adults but this is not an overall phenomenon. The gap between dyslexic and typical readers decreases in adults compared with children in decoding errors, word reading in context and orthographic ability measures; however, it increases in most of the phonological processing tasks (e.g. phonological rhyming). Therefore, studies with children do not clarify which specific deficits persist from childhood even after years of formal schooling and print exposure and the reading domains that are somewhat compensated for in adulthood. In this way, adults with dyslexia might represent a valuable model for studying whether the behavioural manifestations of dyslexia change in the long run.</p> <p>Furthermore, an important question that can be addressed through the study of adults is why do some individuals attain age-appropriate reading skills despite a history of reading and spelling difficulties (compensated or high-functioning dyslexics) while others do not (non-compensated) (Lefly & Pennington, [<reflink idref="bib32" id="ref9">32</reflink>]; for a recent discussion on this topic, see Cavalli et al., [<reflink idref="bib10" id="ref10">10</reflink>]; Eloranta, Närhi, Eklund, Ahonen, & Aro, [<reflink idref="bib16" id="ref11">16</reflink>]). One approach to exploring this question is the cognitive profiling of this population. For instance, Cavalli et al. ([<reflink idref="bib10" id="ref12">10</reflink>]) tried to understand which language abilities dyslexics may rely on to compensate for their deficits. The results revealed the existence of deficits in phonological but not morphological abilities, suggesting that university students with dyslexia may compensate for their reading weaknesses by drawing on morphological knowledge.</p> <p>Understanding the pattern of strengths and weakness of this adult population is critical given the growing number of dyslexic students in higher education institutions (Pino & Mortari, [<reflink idref="bib47" id="ref13">47</reflink>]). To that end, the need to construct adequate assessment and diagnostic protocols has previously been recognised by Callens, Tops, Stevens, and Brysbaert ([<reflink idref="bib8" id="ref14">8</reflink>]). The better and more sensitive the instruments that are used to characterise this specific group, the better the quality of the support that is provided, which means this group will experience fewer difficulties in terms of academic skills and success.</p> <p>Thus, we performed a systematic review of the residual difficulties of adults with dyslexia. The only meta-analytic evidence available so far about adults with reading disorders was obtained by Swanson and Hsieh ([<reflink idref="bib56" id="ref15">56</reflink>]). Adults with reading disorders differ significantly from typical adult readers in measures such as word recognition, pseudoword reading, reading comprehension, spelling, writing, naming speed, phonological processing, verbal memory, vocabulary and verbal intelligence (moderate to high effect sizes). In addition, differences are seen in general cognitive variables, such as problem-solving/reasoning, visual memory, monitoring or executive processing, perceptual skills, general intelligence and personality (low to moderate effect sizes). Furthermore, effect sizes varied as a function of reading and intellectual level; larger effect sizes emerged for studies with relatively high IQs and low overall reading scores. Taken together, these results support the idea that the deficits found in children persist until adulthood.</p> <p>However, with few exceptions, most of the findings concerning adults with dyslexia come from studies conducted in English, which is assumed to have the most opaque alphabetic orthography. In Swanson and Hsieh's meta-analysis, 50 out of 52 studies contained samples that were recruited from English-speaking populations. Hence, the results cannot be generalised to orthographies with different characteristics. The importance of considering the role of orthographic transparency in reading acquisition and dyslexia is empirically justified, given the number of large-scale studies from the last decade showing that the degree of symbol–sound consistency of a language may affect the rate of reading acquisition (Seymour, Aro, & Erskine, [<reflink idref="bib52" id="ref16">52</reflink>]) and the cognitive skills that predict reading performance (Landerl et al., [<reflink idref="bib29" id="ref17">29</reflink>]; Vaessen et al., [<reflink idref="bib62" id="ref18">62</reflink>]; Ziegler et al., [<reflink idref="bib67" id="ref19">67</reflink>]). Moreover, it is well known that reading difficulties are sensitive to orthographic transparency; thus, studies have been conducted to analyse the characteristics of dyslexia in different orthographies (Landerl et al., [<reflink idref="bib30" id="ref20">30</reflink>]; Ziegler, Perry, Ma-Wyatt, Ladner, & Schulte-Körne, [<reflink idref="bib68" id="ref21">68</reflink>]). For instance, in orthographies that are more transparent than English, the difficulties in terms of reading skills are usually determined by reading fluency rather than by accuracy (see, for example, for Italian, Re et al., [<reflink idref="bib50" id="ref22">50</reflink>] and for Spanish, Suárez-Coalla and Cuetos [<reflink idref="bib15" id="ref23">15</reflink>]).</p> <p>As most of the research seeking to characterise dyslexia in adulthood has concentrated on English-based samples, the results cannot be fully generalised to other samples from less opaque orthographies. For instance, Pennington et al. ([<reflink idref="bib44" id="ref24">44</reflink>]) indicated that poor phonological awareness is one of the primary deficits that persist in anglophone adults with dyslexia, while Nergård-Nilssen and Hulme ([<reflink idref="bib37" id="ref25">37</reflink>]) suggested that spelling problems were the most prominent markers of dyslexia in adults in a Norwegian sample (a more transparent orthography). Interestingly, in a Dutch sample (intermediate orthography), Tops, Callens, Lammertyn, van Hees, and Brysbaert ([<reflink idref="bib60" id="ref26">60</reflink>]) identified word reading, word spelling and phoneme reversal time as the tests with the most predictive power to identify dyslexia in adulthood.</p> <p>In light of this heterogeneity, and to fill a gap in Swanson and Hsieh's ([<reflink idref="bib56" id="ref27">56</reflink>]) meta-analysis, we reviewed 178 studies whose distribution in terms of orthographic transparency was as follows: 12.9% samples from transparent orthographies, 19.7% from intermediate orthographies and 67.4% from opaque orthographies. In addition, we extended the previous meta-analytic review by including separate analyses by accuracy and speed measures, as both measures are critical to the diagnosis of dyslexia and potentially interact with orthographic depth (Share, [<reflink idref="bib53" id="ref28">53</reflink>]). This decision was also made based on previous studies showing that speed measures (such as reading fluency) are more suited than accuracy measures to identify reading problems in adults, particularly those coming from more transparent orthographies (Re et al., [<reflink idref="bib50" id="ref29">50</reflink>]; Suárez-Coalla & Cuetos [<reflink idref="bib15" id="ref30">15</reflink>]).</p> <p>Considering the topics currently under discussion regarding dyslexia in adults and the importance of orthographic consistency in the expression of symptoms of dyslexia, we intended in this meta-analytic study to clarify the following critical questions: (<reflink idref="bib1" id="ref31">1</reflink>) To what extent, and in what manner, do adults with reading difficulties (dyslexia) differ from typical adult readers (control group) on measures of overall reading and writing competence and related cognitive skills?; (<reflink idref="bib2" id="ref32">2</reflink>) To what extent do speed measures pose a greater challenge than accuracy measures in an adult population that has already had years of print exposure?; and (<reflink idref="bib3" id="ref33">3</reflink>) To what extent does orthographic transparency modulate the reading profile of adults with dyslexia? The broad age range in the adult sample creates the possibility of analysing the effects of differences in reading practice on the symptoms of dyslexia. Thus, we also analyse whether age moderates the differences between the groups.</p> <hd id="AN0146733574-3">Methods</hd> <p></p> <hd id="AN0146733574-4">Study selection and inclusion criteria</hd> <p>The studies included in the meta-analysis were identified in searches of the PsycINFO and PubMed databases, using a combination of search terms related to reading disorders in adulthood [(<emph>Adult</emph> OR <emph>Students</emph> OR <emph>College Students</emph>) AND (<emph>Dyslexia</emph> OR <emph>Reading Disorders</emph> OR <emph>Reading Disabilities</emph>)] and using, as filters, <emph>Adults</emph> (18 years and older) and <emph>Peer-Reviewed Journals</emph>. The meta-analysis included only studies published in English. Our search covered the title, abstract and keywords of all published articles that were available in the databases from 2006 until November 2019. We choose this time window to prevent an overlap with Swanson and Hsieh's ([<reflink idref="bib56" id="ref34">56</reflink>]) previous meta-analysis. After the removal of duplicates, we found 3196 references. For inclusion, a study had to meet the following criteria: (i) report original empirical data for reading and reading-related variables, (ii) compare the performance of individuals with dyslexia with that of typical control readers (matched by relevant variables), (iii) maintain a sample age over 18 years old and (iv) contain enough information to compute effect sizes. The samples with dyslexia were required to have an <emph>explicit</emph> classification as "dyslexics" or with "specific reading disability/disorder", typically based on a previous diagnosis. To avoid violation of the independence of observations, studies from the same authors were examined for duplicate samples. Whenever this occurred, we included only the articles that had more complete data sets and/or larger samples (see Fig. 1).</p> <p>Graph: Fig. 1 Flow diagram for the search and inclusion of studies</p> <p>In the original search, only 178 articles met all of our inclusion criteria (see References for the studies included in the meta-analysis). Two of the authors independently coded a random sample of 30% of the studies. Intercoder agreement was estimated using the Pearson correlation for continuous information (such as effect sizes) and Cohen's kappa for categorical information (such as type of measure). Intercoder correlation coefficients ranged from 0.96 to 1.0, while the kappa coefficients ranged from 0.85 to 1.0. Disagreements were resolved by consulting the original article or by discussion.</p> <hd id="AN0146733574-5">Coding procedure and moderator variables</hd> <p>For each study, the following information was coded for dyslexic and control (typical readers) groups: sample size, mean age, number of females, mean years of schooling and orthography in which the participants were assessed. Means and standard deviations for the dyslexic and control groups were also extracted for reading and reading-related measures, for accuracy and speed (whenever available). These included the most widely studied measures of core cognitive skills in reading disorders in children and adults and were categorised into three main groups: reading and writing variables (word, pseudoword and text reading, reading comprehension, spelling), reading- and writing-related variables (phonological awareness, orthographic knowledge, phonological memory, rapid automatised naming, verbal working memory and vocabulary) and general cognitive ability (full IQ, verbal and non-verbal IQ, abstraction and speed of processing). Data for all reading and writing measures and reading-related measures were based on tests that were either specifically designed for the study or taken from well-known reading batteries. Examples of tests in these categories are provided.</p> <hd id="AN0146733574-6">Reading and writing variables</hd> <p>The following categories were considered:</p> <p></p> <ulist> <item> <emph>Word reading</emph>: This category included word reading measures aimed at assessing the visual recognition of real words from different categories (e.g. regular and irregular words) by means of single-word identification tasks [e.g. Woodcock Reading Mastery Test (WRMT), Wide Range Achievement Test (WRAT), Test of Word Reading Efficiency (TOWRE), One Minute Test for timed reading of words (OMT) and the Alouette reading Test].</item> <p></p> <item> <emph>Pseudoword reading</emph>: This category included measures of single-pseudoword reading to assess decoding skills [e.g. Wechsler Individual Achievement Test (WIAT), De Klepel Test, Word Attack from Woodcock-Johnson III (WJ3)].</item> <p></p> <item> <emph>Text reading</emph>: This category included measures in which the participant solely has to read a passage from a scholastic book, a text or unconnected sentences, with no need to answer any questions about it [e.g. Gray Oral Reading (GORT), Nelson-Denny Reading Test, Memory-Transfer battery (MT)].</item> <p></p> <item> <emph>Reading comprehension</emph>: This category included measures in which the participant reads a passage or sentence and then has to answer questions about it to assess her/his comprehension [e.g. York Adult Assessment Battery Revised (YAA-R)] or in which the participant is asked to perform a multiple-choice sentence completion task within a time limit (e.g. Reading Age Test; 1-min TIL).</item> <p></p> <item> <emph>Spelling</emph>: This category included measures aimed at assessing the ability to spell using conventions of letter–sound relationships, such as writing words or pseudowords from dictation [e.g. <emph>Outil de DÉpistage des DYSlexies</emph> (ODEDYS), Wechsler Individual Achievement Test (WIAT), Dyslexia Adult Screening Test (DAST)].</item> </ulist> <hd id="AN0146733574-7">Reading- and writing-related variables</hd> <p>The following categories were considered:</p> <p></p> <ulist> <item> <emph>Phonological awareness</emph>: This category included several tasks requiring participants to reflect upon and manipulate the speech sounds of words (e.g. phoneme deletion, spoonerism, syllable reversal and rhyme recognition) [e.g. Comprehensive Test of Phonological Processing (CTOPP), <emph>Batería para la Evaluación de la Competencia Lectora</emph> (EVALEC), YAA-R, DAST, ODEDYS]. Since the assessment of phonological awareness involved tasks with different complexity levels, this category was subdivided into four levels adapted from Parrila et al. ([<reflink idref="bib42" id="ref35">42</reflink>]). Level 1 tasks focus on phonological units larger than phonemes that require simple processing (judgement, matching or segmentation of syllables or rhymes); level 2 tasks focus on phonological units larger than phonemes and require more complex processing (reversal or deletion of syllables and rhymes); level 3 tasks focus on phonemes and require simple phonemic awareness processing (matching, blending and segmentation of phonemes); and level 4 tasks focus on phonemes and require complex manipulations (deletion and substitution of phonemes and spoonerisms).</item> <p></p> <item> <emph>Phonological memory</emph>: This category included tasks in which phonological processing is automatically engaged, such as word and pseudoword repetition (e.g. CTOPP, ODEDYS) and digit span forward tasks (e.g. WAIS).</item> <p></p> <item> <emph>Orthographic knowledge</emph>: This category included tasks in which the participant must access specific information stored in long-term memory about how to represent spoken language in written form, such as word chains, orthographic choice, proofreading, and lexical decision tasks [e.g. Dyslexia Screening Battery (DUVAN)].</item> <p></p> <item> <emph>Verbal working memory</emph>: This category included the classic digit span backward measure from batteries such as the Wechsler Adult Intelligence Scale (WAIS).</item> <p></p> <item> <emph>Vocabulary</emph>: This category included measures of semantic knowledge about words and objects [e.g. vocabulary test from WAIS, Peabody Picture Vocabulary Test (PPVT-4), Shipley vocabulary and <emph>Échelle de Vocabulaire en Images Peabody</emph> (EVIP)]. Given that vocabulary measures are occasionally used as matching criteria for selecting participants, only effect sizes (ESs) from studies in which vocabulary was exclusively used to characterise groups were included to avoid underestimating the true vocabulary deficits.</item> <p></p> <item> <emph>Rapid automatised naming (RAN)</emph>: This category included measures of phonological retrieval and processing, assessed through tasks featuring a continuous matrix presentation of alphabetic (letters and digits) and non-alphabetic (objects and colours) stimuli, as in the standard serial Rapid Automatized Naming Test (Denckla & Rudel, [<reflink idref="bib14" id="ref36">14</reflink>]).</item> </ulist> <hd id="AN0146733574-8">General cognitive variables</hd> <p>The following categories were considered:</p> <p></p> <ulist> <item> <emph>Intelligence (IQ)</emph>: The intelligence measures were coded to distinguish between full IQ (e.g. WAIS full scale), verbal IQ (e.g. Woodcock-Johnson III—WJ-III-COG and the verbal scale from WAIS) and non-verbal IQ (e.g. Raven Progressive Matrices and the performance scale from WAIS). In order to not underestimate the true differences between groups, only ESs from studies in which IQ measures were not used as criteria for matching groups were included in the analysis.</item> <p></p> <item> <emph>Abstraction</emph>: This category included the similarities subtest from WAIS.</item> <p></p> <item> <emph>Speed of processing</emph>: This category included tasks with a speed factor (e.g. Digit-Symbol Coding from WAIS and reaction time tasks).</item> </ulist> <p>We note that other cognitive variables (e.g. non-verbal working memory, attention, oral comprehension, writing and calculation) were also considered but were not included in the analyses because there were only a few studies (< 7) for each.</p> <hd id="AN0146733574-9">Task measures</hd> <p>Task scores were classified as accuracy or speed measures. A score was considered as an accuracy measure when it corresponded to the number, percentage and/or proportion of correct responses/errors in a time-unlimited task. A score was classified as a speed measure when the performance was based on the number of correct responses by time-limited tasks (correct responses per second) and when the score corresponded to the time spent completing a time-unlimited task.</p> <hd id="AN0146733574-10">Orthography</hd> <p>Orthographic consistency was coded into three categories of complexity expressing the feedforward and feedback consistency of grapheme–phoneme and phoneme–grapheme correspondences (like a recent large-scale cross-linguistic study by Landerl et al., [<reflink idref="bib30" id="ref37">30</reflink>]) and based on Seymour, Aro, and Erskine's ([<reflink idref="bib52" id="ref38">52</reflink>]) classification: "opaque" (Danish, English, French and Hebrew) and "transparent" (Spanish, Icelandic, Norwegian, Italian, Finnish, Polish), with high and low levels of inconsistencies in both directions, respectively, and "intermediate" complexity level (Dutch, German, European-Portuguese and Swedish) with high feedforward consistency but low feedback consistency or the reverse. Our classification was similar to the one used in a recent meta-analysis (Araújo & Faísca, [<reflink idref="bib1" id="ref39">1</reflink>]). Also, based on Verhoeven and Perfetti ([<reflink idref="bib64" id="ref40">64</reflink>]), consonantal root-based writing systems such as Hebrew were included in the opaque category.</p> <hd id="AN0146733574-11">Meta-analytic procedures</hd> <p></p> <hd id="AN0146733574-12">Effect size estimates</hd> <p>Data were analysed using the Comprehensive Meta-Analysis software (Borenstein, Hedges, Higgins, & Rothstein, [<reflink idref="bib4" id="ref41">4</reflink>]). Effect sizes involving group comparisons between readers with dyslexia and typical control-matched readers were computed with Cohen's <emph>d</emph> with corrections for small sample sizes (Hedges, [<reflink idref="bib25" id="ref42">25</reflink>]). Effect size estimates were, in some cases (e.g. a study reporting two measures of real word reading), aggregated using the arithmetic mean to avoid over-representation of multi-experiment studies in the overall analyses (Rosenthal, [<reflink idref="bib51" id="ref43">51</reflink>]). When means and standard deviations were not provided, <emph>d</emph> values were estimated from the reported <emph>t</emph> or <emph>F</emph> statistics. A positive <emph>d</emph> value indicated that the control subjects had the highest group mean. A commonly used interpretation is to refer to effect sizes as small (<emph>d</emph> = 0.2), medium (<emph>d</emph> = 0.5) and large (<emph>d</emph> = 0.8) based on Cohen's ([<reflink idref="bib12" id="ref44">12</reflink>]) recommended guidelines.</p> <p>When individual studies reported more than one effect size of interest for the same sample, we used the shifting unit of analysis approach (Cooper, [<reflink idref="bib13" id="ref45">13</reflink>]), as this procedure provides a good compromise between preserving the independence of the effect sizes and retaining the maximum amount of information from each study. In this approach, each effect size associated with a sample is first coded as if it was an independent estimate of the relationship. The unit of analysis is then shifted according to the hypothesis being tested. For the overall mean effect analysis, and whenever the moderator corresponded to a between-studies factor defining separate groups of participants (e.g. orthography), we used the sample as a unit of analysis. The multiple effects from each sample were aggregated so that each sample contributed only one effect and all sample averaged effects were (almost) independent. When testing the moderator effect of an outcome domain (e.g. type of reading measure), and when multiple effect sizes were available within the same sample, we shifted the unit of analysis from the sample to the effect sizes, thereby allowing each sample to contribute one effect size to each category of the moderator.</p> <hd id="AN0146733574-13">Analysis of effect sizes</hd> <p>Overall effect sizes were estimated by calculating a weighted average of individual effect sizes (Rosenthal, [<reflink idref="bib51" id="ref46">51</reflink>]) based on a random effects model. For each meta-analysis, we calculated a 95% confidence interval (CI), <emph>Z</emph> significance test for null effects and its <emph>p</emph> value, within-group heterogeneity (<emph>Q</emph><subs>Within</subs>) and the percentage of variation across studies due to heterogeneity rather than sampling error (<emph>I</emph><sups>2</sups>) (Borenstein, Hedges, Higgins, & Rothstein, [<reflink idref="bib5" id="ref47">5</reflink>]).</p> <p>To test the categorical moderator effects, we proceeded with a subgroup analysis with the mixed-effects between-groups heterogeneity statistics (<emph>Q</emph><subs>Between</subs>, which has a chi-square distribution and is analogous to an analysis of variance <emph>F</emph> test). For the continuous variables, we used a meta-regression based on the method of moments for the random effects model to predict variations in effect size across studies from the moderator variables. We report the percentage of between-study variance explained (<emph>R</emph><sups>2</sups>) as a measure of the effect size of the moderator.</p> <p>Forest plots were used to examine the distributions of effect sizes and to detect potential outliers, while sensitivity analyses were conducted to determine their impact on the overall range of means. Sensitivity analyses allow an adjusted overall effect size to be estimated after the removal of studies one by one. Finally, we examined funnel plots for random effects models to determine the presence of publication bias. The "trim and fill" method for random effects models (Duval & Tweedie, [<reflink idref="bib15" id="ref48">15</reflink>]) was used to examine the impact of possible missing studies.</p> <hd id="AN0146733574-14">Results</hd> <p>The present meta-analytic study included 178 articles (185 samples) comprising 1817 ESs comparing adults with dyslexia with matched typical readers. The total sample size was 4363 (mean sample size = 23.6; SD = 17.4) for adults with dyslexia and 5029 (<emph>M</emph> = 27.2; SD = 23.9) for the control group. Concerning the reading status of the dyslexic participants, 151 studies (84.8%) explicitly stated that dyslexic individuals had a previous formal diagnosis of dyslexia; 21 studies (11.8%) included participants who self-reported a history of reading problems, while in six studies (3.4%), the reading disorder group was selected from a large pool of participants (in both cases, the authors confirmed the participants' reading status with an appropriate battery). In addition, information about attentional problems was reported for about half of the studies: 103 studies (57.9%) explicitly claimed that participants with attentional problems were excluded, while three studies (1.7%) included, in their samples, some participants with or at risk of attentional problems and/or hyperactivity. Given the low representativeness of these participants in the samples, we decided to keep these three studies in order to avoid the loss of data. For 72 studies (40.4%), the authors did not mention whether their samples included or did not include participants with attention problems.</p> <p>Overall, the dyslexic and control groups were equivalent in age with a weighted mean age of 24.7 years (<emph>k</emph> = 148 samples; SEM = 0.80; 95% CI [23.1, 26.4]) and 24.2 years (<emph>k</emph> = 148 samples; SEM = 0.74; 95% CI [22.7, 25.7]), respectively. The sex ratio was also equivalent between samples (<emph>k</emph> = 151 samples; female % in dyslexic samples: mean = 55.1%; 95% CI [52.4, 57.7]; female % in control samples: mean = 57.6%; 95% CI [54.9, 60.3]). Thirty studies did not report information about the participants' educational level, 100 included samples that were recruited at universities, 11 included samples with varied educational levels (conjugating basic, middle, high educational levels and university students) and only 37 reported the average years of schooling (<emph>k</emph> = 39 dyslexic samples: mean = 14.0 years; SEM = 0.18; 95% CI [13.7, 14.4]; <emph>k</emph> = 39 control samples: mean = 13.9 years; SEM = 0.17; 95% CI [13.6, 14.3]).</p> <p>To fulfil our goals, the data was next analysed and reported based on our three main questions. Additionally, we used a meta-regression analysis to analyse whether the participants' age modulated reading deficits. Whenever the group comparisons included fewer than six effect sizes for a given category, the results were not reported. Positive ES favoured adults without reading disorders.</p> <p> <emph>To what extent, and in what manner, do adults with reading difficulties (dyslexia) differ from typical adult readers (control group) on measures of overall reading and writing competence and related cognitive skills?</emph> </p> <p>Overall, adult readers with dyslexia show poor performance compared with controls on all reading and writing tasks, with very large deficits (<emph>d</emph> > 1.7) in all measures except reading comprehension (<emph>d</emph> = 0.729). They also struggle to a lesser extent (0.5 < <emph>d</emph> < 1.3) with tasks that place demands on their phonological awareness and memory, orthographic knowledge and rapid automatised naming skills. Concerning general cognitive skills, as expected, the observed ESs vary from small (abstraction: <emph>d</emph> = 0.143; non-verbal IQ: <emph>d</emph> = 0.187) to large (speed of processing: <emph>d</emph> = 0.840) with some ESs in the small-to-medium range (full IQ: <emph>d</emph> = 0.296; verbal IQ: <emph>d</emph> = 0.424). For a clearer characterisation of adults' reading difficulties, each skill will be analysed separately (see Table 1).</p> <p>Effect sizes with 95% confidence intervals and heterogeneity statistics in studies comparing adults with dyslexia with typical readers</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th rowspan="2" /><th rowspan="2"><p><italic>k</italic></p></th><th colspan="2"><p>Effect size</p></th><th rowspan="2"><p><italic>Z</italic></p></th><th rowspan="2"><p><italic>p</italic> value</p></th><th colspan="3"><p>Heterogeneity</p></th></tr><tr><th><p><italic>d</italic></p></th><th><p>95% CI</p></th><th><p><italic>I</italic><sup>2</sup></p></th><th><p><italic>Q</italic><sub>Within</sub></p></th><th><p><italic>p</italic> value</p></th></tr></thead><tbody><tr><td colspan="9"><p>Reading and writing variables</p></td></tr><tr><td><p> Word read</p></td><td><p>161</p></td><td><p>1.812</p></td><td><p>1.690–1.935</p></td><td><p>29.0</p></td><td><p>< 0.001</p></td><td><p>79.43</p></td><td><p>777.7</p></td><td><p>< 0.001</p></td></tr><tr><td><p> Pseudoword read</p></td><td><p>136</p></td><td><p>2.034</p></td><td><p>1.896–2.172</p></td><td><p>28.8</p></td><td><p>< 0.001</p></td><td><p>80.61</p></td><td><p>796.1</p></td><td><p>< 0.001</p></td></tr><tr><td><p> Text read</p></td><td><p>39</p></td><td><p>1.761</p></td><td><p>1.472–2.050</p></td><td><p>11.4</p></td><td><p>< 0.001</p></td><td><p>86.41</p></td><td><p>279.6</p></td><td><p>< 0.001</p></td></tr><tr><td><p> Reading comprehension</p></td><td><p>37</p></td><td><p>0.729</p></td><td><p>0.550–0.907</p></td><td><p>8.0</p></td><td><p>< 0.001</p></td><td><p>74.48</p></td><td><p>141.1</p></td><td><p>< 0.001</p></td></tr><tr><td><p> Spelling</p></td><td><p>99</p></td><td><p>1.735</p></td><td><p>1.590–1.880</p></td><td><p>23.5</p></td><td><p>< 0.001</p></td><td><p>78.06</p></td><td><p>446.7</p></td><td><p>< 0.001</p></td></tr><tr><td colspan="9"><p>Reading- and writing-related variables</p></td></tr><tr><td><p> Phonological awareness</p></td><td><p>101</p></td><td><p>1.177</p></td><td><p>1.075–1.279</p></td><td><p>22.7</p></td><td><p>< 0.001</p></td><td><p>60.02</p></td><td><p>250.2</p></td><td><p>< 0.001</p></td></tr><tr><td><p> Level 1</p></td><td><p>2</p></td><td><p>0.757</p></td><td><p>0.117–1.387</p></td><td><p>2.3</p></td><td><p>0.020</p></td><td><p>6.41</p></td><td><p>1.1</p></td><td><p>0.301</p></td></tr><tr><td><p> Level 2</p></td><td><p>6</p></td><td><p>1.290</p></td><td><p>0.849–1.731</p></td><td><p>5.7</p></td><td><p>< 0.001</p></td><td><p>53.14</p></td><td><p>10.7</p></td><td><p>0.054</p></td></tr><tr><td><p> Level 3</p></td><td><p>21</p></td><td><p>0.779</p></td><td><p>0.563–0.995</p></td><td><p>7.1</p></td><td><p>< 0.001</p></td><td><p>53.50</p></td><td><p>43.0</p></td><td><p>0.002</p></td></tr><tr><td><p> Level 4</p></td><td><p>85</p></td><td><p>1.263</p></td><td><p>1.156–1.371</p></td><td><p>23.1</p></td><td><p>< 0.001</p></td><td><p>56.88</p></td><td><p>194.8</p></td><td><p>< 0.001</p></td></tr><tr><td><p> Phonological memory</p></td><td><p>66</p></td><td><p>1.034</p></td><td><p>0.916–1.153</p></td><td><p>17.1</p></td><td><p>< 0.001</p></td><td><p>57.99</p></td><td><p>154.7</p></td><td><p>< 0.001</p></td></tr><tr><td><p> Orthographic knowledge</p></td><td><p>26</p></td><td><p>1.233</p></td><td><p>1.043–1.423</p></td><td><p>12.7</p></td><td><p>< 0.001</p></td><td><p>65.59</p></td><td><p>72.7</p></td><td><p>< 0.001</p></td></tr><tr><td><p> Verbal working memory</p></td><td><p>26</p></td><td><p>0.926</p></td><td><p>0.694–1.158</p></td><td><p>7.8</p></td><td><p>< 0.001</p></td><td><p>68.64</p></td><td><p>79.7</p></td><td><p>< 0.001</p></td></tr><tr><td><p> Vocabulary</p></td><td><p>38</p></td><td><p>0.591</p></td><td><p>0.440–0.742</p></td><td><p>7.7</p></td><td><p>< 0.001</p></td><td><p>59.34</p></td><td><p>91.0</p></td><td><p>< 0.001</p></td></tr><tr><td><p> RAN</p></td><td><p>77</p></td><td><p>1.191</p></td><td><p>1.091–1.290</p></td><td><p>23.4</p></td><td><p>< 0.001</p></td><td><p>45.07</p></td><td><p>138.4</p></td><td><p>< 0.001</p></td></tr><tr><td><p> RAN alphabetic</p></td><td><p>57</p></td><td><p>1.295</p></td><td><p>1.154–1.437</p></td><td><p>18.0</p></td><td><p>< 0.001</p></td><td><p>63.30</p></td><td><p>152.6</p></td><td><p>< 0.001</p></td></tr><tr><td><p> RAN non-alphabetic</p></td><td><p>50</p></td><td><p>0.972</p></td><td><p>0.816–1.128</p></td><td><p>12.2</p></td><td><p>< 0.001</p></td><td><p>68.62</p></td><td><p>156.2</p></td><td><p>< 0.001</p></td></tr><tr><td colspan="9"><p>General cognitive skills</p></td></tr><tr><td><p> Full IQ</p></td><td><p>28</p></td><td><p>0.296</p></td><td><p>0.158–0.434</p></td><td><p>4.2</p></td><td><p>< 0.001</p></td><td><p>44.58</p></td><td><p>48.7</p></td><td><p>0.006</p></td></tr><tr><td><p> Non-verbal IQ</p></td><td><p>75</p></td><td><p>0.187</p></td><td><p>0.106–0.269</p></td><td><p>4.5</p></td><td><p>< 0.001</p></td><td><p>23.82</p></td><td><p>97.1</p></td><td><p>0.037</p></td></tr><tr><td><p> Verbal IQ</p></td><td><p>10</p></td><td><p>0.424</p></td><td><p>0.028–0.821</p></td><td><p>2.1</p></td><td><p>0.036</p></td><td><p>75.27</p></td><td><p>36.4</p></td><td><p>< 0.001</p></td></tr><tr><td><p> Abstraction</p></td><td><p>13</p></td><td><p>0.143</p></td><td><p>− 0.121 to 0.408</p></td><td><p>1.1</p></td><td><p>0.287</p></td><td><p>59.80</p></td><td><p>29.8</p></td><td><p>0.003</p></td></tr><tr><td><p> Speed of processing</p></td><td><p>11</p></td><td><p>0.840</p></td><td><p>0.551–1.129</p></td><td><p>5.7</p></td><td><p>< 0.001</p></td><td><p>72.46</p></td><td><p>36.3</p></td><td><p>< 0.001</p></td></tr></tbody></table> </ephtml> </p> <p>Number of independent effect sizes (<emph>k</emph>) that contributed to each meta-analysis; the weighted average effect size of group differences (<emph>d</emph>); 95% confidence interval (CI); <emph>Z</emph> significance test statistic for the effect sizes and corresponding <emph>p value</emph>; the proportion of total variance between effect sizes not explained by chance (<emph>I</emph><sups>2</sups>); within-group heterogeneity of variance (<emph>Q</emph><subs>Within</subs>) and corresponding <emph>p value</emph></p> <hd id="AN0146733574-15">Word Reading</hd> <p>One hundred and sixty-one independent comparisons were made for word reading in adults with dyslexia and controls. The overall mean effect size was large and significant (<emph>d</emph> = 1.812, 95% CI [1.690, 1.935], <emph>p</emph> < 0.001), confirming that adults with dyslexia perform much more poorly on isolated word reading than do controls without reading difficulties. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 1.786 (95% CI [1.667, 1.904]) to 1.821 (95% CI [1.700, 1.943]). In a trim and fill analysis, 12 studies were imputed to the right of the mean (the adjusted overall mean was 1.927 (95% CI [1.795, 2.059]). The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib160" id="ref49">160</reflink>) = 777.7, <emph>p</emph> < 0.001, and around 80% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 79.43).</p> <hd id="AN0146733574-16">Pseudoword reading</hd> <p>One hundred and thirty-six independent effect sizes compared pseudoword reading performance in adults with dyslexia and controls. The overall mean effect size was large and significant (<emph>d</emph> = 2.034, 95% CI [1.896, 2.172], <emph>p</emph> < 0.001), confirming that adults with dyslexia perform much worse on isolated pseudoword reading than do controls without reading difficulties. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 2.012 (95% CI [1.878, 2.145]) to 2.046 (95% CI [1.909, 2.182]). In a trim and fill analysis, nine studies were imputed to the right of the mean (the adjusted overall mean was 2.136 (95% CI [1.988, 2.285]). The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib135" id="ref50">135</reflink>) = 796.1, <emph>p</emph> < 0.001, and around 80% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 80.61).</p> <hd id="AN0146733574-17">Text reading</hd> <p>Thirty-nine independent effect sizes were computed for text reading measures. The overall mean effect size was large and significant (<emph>d</emph> = 1.761, 95% CI [1.472, 2.050], <emph>p</emph> < 0.001), confirming that adults with dyslexia perform much more poorly than controls when context assists reading. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 1.657 (95% CI [1.403, 1.911]) to 1.761 (95% CI [1.472, 2.050]). In a trim and fill analysis, six studies were imputed to the right of the mean (the adjusted overall mean was 1.980 (95% CI [1.665, 2.295]). The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib38" id="ref51">38</reflink>) = 279.6, <emph>p</emph> < 0.001, and over 85% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 86.41).</p> <hd id="AN0146733574-18">Reading comprehension</hd> <p>Thirty-seven independent effect sizes were computed for reading comprehension measures. The overall mean effect size was close to large and significant (<emph>d</emph> = 0.729, 95% CI [0.550, 0.907], <emph>p</emph> < 0.001), confirming that adults with dyslexia perform more poorly on reading comprehension as compared with controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 0.699 (95% CI [0.524, 0.874]) to 0.764 (95% CI [0.550, 0.907]). In a trim and fill analysis, no study was imputed, suggesting the absence of publication bias. The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib36" id="ref52">36</reflink>) = 141.1, <emph>p</emph> < 0.001, and around 75% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 74.48).</p> <hd id="AN0146733574-19">Spelling</hd> <p>Ninety-nine independent effect sizes compared spelling performance in adults with dyslexia and in controls. The overall mean effect size was large and significant (<emph>d</emph> = 1.735, 95% CI [1.590, 1.880], <emph>p</emph> < 0.001), confirming that adults with dyslexia perform much worse on spelling measures than do controls without reading difficulties. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 1.717 (95% CI [1.574, 1.860]) to 1.750 (95% CI [1.607, 1.893]). There was evidence of publication bias favouring studies with larger effect sizes, confirmed by the funnel plot and the trim and fill analysis (23 studies imputed were on the left side of the mean; the adjusted overall mean was 1.431 and 95% CI [1.369, 1.493]). The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib98" id="ref53">98</reflink>) = 446.7, <emph>p</emph> < 0.001, and almost 80% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 78.06).</p> <hd id="AN0146733574-20">Phonological awareness</hd> <p>One hundred and one independent effect sizes were computed for the phonological awareness measures. The overall mean effect size was large and significant (<emph>d</emph> = 1.177, 95% CI [1.075, 1.279], <emph>p</emph> < 0.001), confirming that adults with dyslexia perform more poorly on phonological awareness measures than do controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 1.161 (95% CI [1.062, 1.260]) to 1.188 (95% CI [1.089, 1.287]). There was evidence of publication bias favouring studies with larger effect sizes, confirmed by the funnel plot and the trim and fill analysis (31 studies imputed on the left side of the mean; the adjusted overall mean was 0.934 and 95% CI [0.821, 1.045]). The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib100" id="ref54">100</reflink>) = 250.2, <emph>p</emph> < 0.001, and 60% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 60.02). Beyond the global effect, Table 1 also presents the ESs disaggregated by the level of processing complexity required by the phonological awareness tasks. There was a significant effect of levels on the magnitude of the phonological awareness deficit (<emph>p</emph> < 0.001; level 1 was excluded from the analysis due to the number of ESs). Level 3 presented a large ES, though still significantly smaller than the other two levels (level 2 vs. level 3, <emph>p</emph> = 0.040; level 2 vs. level 4, <emph>p</emph> = 0.911; level 3 vs. level 4, <emph>p</emph> < 0.001).</p> <hd id="AN0146733574-21">Phonological memory</hd> <p>Sixty-six independent effect sizes were computed for phonological memory measures. The overall mean effect size was large and significant (<emph>d</emph> = 1.034, 95% CI [0. 916, 1.153], <emph>p</emph> < 0.001), confirming that adults with dyslexia perform more poorly on phonological memory measures as compared with controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 0.995 (95% CI [0.894, 1.097]) to 1.050 (95% CI [0.934, 1.166]). In a trim and fill analysis, no study was imputed, suggesting the absence of publication bias. The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib65" id="ref55">65</reflink>) = 154.7, <emph>p</emph> < 0.001, and around 60% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 57.99).</p> <hd id="AN0146733574-22">Orthographic knowledge</hd> <p>Twenty-six independent effect sizes were computed for orthographic knowledge measures. The overall mean effect size was large and significant (<emph>d</emph> = 1.233, 95% CI [1.043, 1.423], <emph>p</emph> < 0.001), confirming that adults with dyslexia perform more poorly on tasks involving orthographic knowledge as compared with controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 1.199 (95% CI [1.012, 1.386]) to 1.273 (95% CI [1.092, 1.453]). There is evidence of publication bias favouring studies with larger effect sizes, confirmed by the funnel plot and the trim and fill analysis (one study imputed on the left side of the mean; the adjusted overall mean was 1.210 and 95% CI [1.020, 1.399]). The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib25" id="ref56">25</reflink>) = 72.7, <emph>p</emph> < 0.001, and over 60% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 65.59).</p> <hd id="AN0146733574-23">Verbal working memory</hd> <p>Twenty-six independent effect sizes were computed for verbal working memory. The overall mean effect size was large and significant (<emph>d</emph> = 0.926, 95% CI [0.694, 1.158], <emph>p</emph> < 0.001), confirming that adults with dyslexia perform more poorly on verbal working memory tasks as compared with controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 0.810 (95% CI [0.684, 0.936]) to 0.950 (95% CI [0.712, 1.189]). In a trim and fill analysis, nine studies were imputed to the right of the mean (the adjusted overall mean was 1.181 (95% CI [0.953, 1.409]). The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib25" id="ref57">25</reflink>) = 79.7, <emph>p</emph> < 0.001, and almost 70% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 68.64).</p> <hd id="AN0146733574-24">Vocabulary</hd> <p>Thirty-eight independent effect sizes were computed for vocabulary measures. The overall mean effect size was medium and significant (<emph>d</emph> = 0.591, 95% CI [0.440, 0.742], <emph>p</emph> < 0.001), suggesting that adults with dyslexia have weaker vocabulary skills than do controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 0.562 (95% CI [0.417, 0.707]) to 0.615 (95% CI [0.467, 0.763]). In a trim and fill analysis, three studies were imputed to the right of the mean (the adjusted overall mean was 0.652 (95% CI [0.497, 0.806]). This was confirmed by the funnel plot. The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib37" id="ref58">37</reflink>) = 91.0, <emph>p</emph> < 0.001, and almost 60% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 59.34).</p> <hd id="AN0146733574-25">Rapid automatised naming</hd> <p>Seventy-seven independent effect sizes were computed for rapid automatised naming. The overall mean effect size was large and significant (<emph>d</emph> = 1.191, 95% CI [1.091, 1.290], <emph>p</emph> < 0.001), confirming that adults with dyslexia perform more poorly on rapid automatised naming tasks than do controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 1.175 (95% CI [1.078, 1.273]) to 1.207 (95% CI [1.116, 1.298]). There was evidence of publication bias, confirmed by the funnel plot and the trim and fill analysis (23 studies imputed on the left of the mean; the adjusted overall mean was 0.980 and 95% CI [0.870, 1.090]). The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib76" id="ref59">76</reflink>) = 138.4, <emph>p</emph> < 0.001, and less than 50% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 45.07). Part of this heterogeneity seems to be explained by the alphabetic nature of the stimulus (see Table 1): the ES associated with RAN alphabetic (letters and digits, <emph>d</emph> = 1.295) is significantly larger (<emph>p</emph> < 0.001) than the ES of RAN non-alphabetic (objects and colours, <emph>d</emph> = 0.972).</p> <hd id="AN0146733574-26">Full IQ</hd> <p>Twenty-eight independent effect sizes were computed for full IQ. The overall mean effect size was significant (<emph>d</emph> = 0.296, 95% CI [0.158, 0.434], <emph>p</emph> < 0.001), confirming that adults with dyslexia have a somewhat lower performance on full IQ measures as compared with controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 0.258 (95% CI [0.132, 0.383]) to 0.320 (95% CI [0.184, 0.455]). There was no evidence of publication bias, confirmed by the funnel plot and the trim and fill analysis (0 studies imputed). The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib27" id="ref60">27</reflink>) = 48.7, <emph>p</emph> = 0.006, and nearly 45% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 44.58).</p> <hd id="AN0146733574-27">Non-verbal IQ</hd> <p>Seventy-five independent effect sizes were computed for non-verbal IQ. The overall mean effect size was small but significant (<emph>d</emph> = 0.187, 95% CI [0.106, 0.269], <emph>p</emph> < 0.001). A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 0.174 (95% CI [0.093, 0.254]) to 0.199 (95% CI [0.119, 0.278]). There was evidence of publication bias, confirmed by the funnel plot and the trim and fill analysis (11 studies imputed on the left of the mean; the adjusted overall mean was 0.108 and 95% CI [0.019, 0.198]). The heterogeneity test was not significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib74" id="ref61">74</reflink>) = 97.1, <emph>p</emph> = 0.037, and nearly 25% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 23.82).</p> <hd id="AN0146733574-28">Verbal IQ</hd> <p>Ten independent effect sizes were computed for verbal IQ. The overall mean effect size was small to medium and significant (<emph>d</emph> = 0.424, 95% CI [0.028, 0.821], <emph>p</emph> = 0.036), confirming that adults with dyslexia have a lower verbal IQ as compared with controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 0.320 (95% CI [− 0.056, 0.696]) to 0.556 (95% CI [0.232, 0.881]). There was no evidence of publication bias, confirmed by the funnel plot and the trim and fill analysis. The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib9" id="ref62">9</reflink>) = 36.4, <emph>p</emph> < 0.001, and three-quarters of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 75.27).</p> <hd id="AN0146733574-29">Abstraction</hd> <p>Thirteen independent effect sizes were computed for abstraction. The overall mean effect size was small (<emph>d</emph> = 0.143, 95% CI [− 0.121, 0.408], <emph>p</emph> = 0.287), suggesting almost equivalent abstraction skills between adults with dyslexia and controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 0.020 (95% CI [− 0.175, 0.216]) to 0.200 (95% CI [− 0.066, 0.467]). There was no evidence of publication bias. The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib13" id="ref63">13</reflink>) = 29.8, <emph>p</emph> = 0.003, and almost 60% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 59.8).</p> <hd id="AN0146733574-30">Speed of processing</hd> <p>Eleven independent effect sizes were computed for speed of processing. The overall mean effect size was large and significant (<emph>d</emph> = 0.840, 95% CI [0.551, 1.129], <emph>p</emph> < 0.001), confirming that adults with dyslexia have a lower speed of processing as compared with controls. A sensitivity analysis showed that after the removal of potential outliers, the overall effect size was in the range of 0.778 (95% CI [0.491, 1.026]) to 0.906 (95% CI [0.613, 1.199]). There was no evidence of publication bias, confirmed by the funnel plot and the trim and fill analysis. The heterogeneity test was significant, <emph>Q</emph><subs>Within</subs> (<reflink idref="bib10" id="ref64">10</reflink>) = 36.3, <emph>p</emph> < 0.001, and over 70% of the observed variance was not accounted for by sampling error alone (<emph>I</emph><sups>2</sups> = 72.46).</p> <p> <emph>To what extent do speed measures pose a greater challenge than accuracy measures in an adult population that has already had years of print exposure?</emph> </p> <p>To address this question, we compared the performance of both groups on accuracy- and speed-based measures of assessment (see Table 2). It is interesting to observe that speed measures exacerbate the reading symptoms of dyslexia. The ESs for real word (<emph>d</emph> = 1.914), pseudoword (<emph>d</emph> = 2.086), text reading (<emph>d</emph> = 1.767), reading comprehension (<emph>d</emph> = 0.900) and spelling (<emph>d</emph> = 1.767) were larger when performance was measured with time. However, the ANOVA analogue analysis indicated significant results only for real word (<emph>p</emph> < 0.001) and pseudoword reading (<emph>p</emph> < 0.001). For reading comprehension, text reading and spelling, the differences between speed and accuracy measures were not significant.</p> <p>Effect sizes with 95% confidence intervals and heterogeneity between measures (accuracy vs. speed) in studies comparing adults with dyslexia with typical readers</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th rowspan="2" /><th rowspan="2"><p><italic>k</italic></p></th><th colspan="2"><p>Effect size</p></th><th rowspan="2"><p><italic>Z</italic></p></th><th rowspan="2"><p><italic>p</italic> value</p></th><th colspan="3"><p>Heterogeneity</p></th></tr><tr><th><p><italic>d</italic></p></th><th><p>95% CI</p></th><th><p><italic>I</italic><sup>2</sup></p></th><th><p><italic>Q</italic><sub>Between</sub></p></th><th><p><italic>p</italic> value</p></th></tr></thead><tbody><tr><td colspan="9"><p>Reading and writing variables</p></td></tr><tr><td colspan="9"><p> Word read</p></td></tr><tr><td><p> Accuracy</p></td><td><p>106</p></td><td><p>1.521</p></td><td><p>1.377–1.665</p></td><td><p>20.7</p></td><td><p>< 0.001</p></td><td><p>78.31</p></td><td char="." align="char"><p>14.57</p></td><td char="." align="char"><p>< 0.001</p></td></tr><tr><td><p> Speed</p></td><td><p>130</p></td><td><p>1.914</p></td><td><p>1.733–2.056</p></td><td><p>26.8</p></td><td><p>< 0.001</p></td><td><p>81.01</p></td><td /><td /></tr><tr><td colspan="9"><p> Pseudoword read</p></td></tr><tr><td><p> Accuracy</p></td><td><p>84</p></td><td><p>1.703</p></td><td><p>1.540–1.867</p></td><td><p>20.4</p></td><td><p>< 0.001</p></td><td><p>77.30</p></td><td char="." align="char"><p>11.24</p></td><td char="." align="char"><p>< 0.001</p></td></tr><tr><td><p> Speed</p></td><td><p>117</p></td><td><p>2.086</p></td><td><p>1.933–2.239</p></td><td><p>26.8</p></td><td><p>< 0.001</p></td><td><p>82.07</p></td><td /><td /></tr><tr><td colspan="9"><p> Text read</p></td></tr><tr><td><p> Accuracy</p></td><td><p>21</p></td><td><p>1.629</p></td><td><p>1.253–2.004</p></td><td><p>8.5</p></td><td><p>< 0.001</p></td><td><p>87.35</p></td><td char="." align="char"><p>0.30</p></td><td char="." align="char"><p>0.584</p></td></tr><tr><td><p> Speed</p></td><td><p>35</p></td><td><p>1.767</p></td><td><p>1.445–2.088</p></td><td><p>10.8</p></td><td><p>< 0.001</p></td><td><p>88.07</p></td><td /><td /></tr><tr><td colspan="9"><p> Reading comprehension</p></td></tr><tr><td><p> Accuracy</p></td><td><p>36</p></td><td><p>0.703</p></td><td><p>0.520–0.887</p></td><td><p>7.5</p></td><td><p>< 0.001</p></td><td><p>74.96</p></td><td char="." align="char"><p>0.57</p></td><td char="." align="char"><p>0.452</p></td></tr><tr><td><p> Speed</p></td><td><p>6</p></td><td><p>0.900</p></td><td><p>0.422–1.378</p></td><td><p>3.7</p></td><td><p>< 0.001</p></td><td><p>75.67</p></td><td /><td /></tr><tr><td colspan="9"><p> Spelling</p></td></tr><tr><td><p> Accuracy</p></td><td><p>96</p></td><td><p>1.720</p></td><td><p>1.573–1.868</p></td><td><p>22.9</p></td><td><p>< 0.001</p></td><td><p>78.59</p></td><td char="." align="char"><p>0.03</p></td><td char="." align="char"><p>0.862</p></td></tr><tr><td><p> Speed</p></td><td><p>9</p></td><td><p>1.767</p></td><td><p>1.261–2.273</p></td><td><p>6.9</p></td><td><p>< 0.001</p></td><td><p>69.21</p></td><td /><td /></tr><tr><td colspan="9"><p>Reading- and writing-related variables</p></td></tr><tr><td colspan="9"><p> Phonological awareness</p></td></tr><tr><td><p> Accuracy</p></td><td><p>94</p></td><td><p>1.074</p></td><td><p>0.964–1.183</p></td><td><p>19.2</p></td><td><p>< 0.001</p></td><td><p>62.79</p></td><td char="." align="char"><p>11.34</p></td><td char="." align="char"><p>< 0.001</p></td></tr><tr><td><p> Speed</p></td><td><p>46</p></td><td><p>1.404</p></td><td><p>1.246–1.562</p></td><td><p>17.4</p></td><td><p>< 0.001</p></td><td><p>64.80</p></td><td /><td /></tr><tr><td colspan="9"><p> Orthographic knowledge</p></td></tr><tr><td><p> Accuracy</p></td><td><p>21</p></td><td><p>1.027</p></td><td><p>0.857–1.217</p></td><td><p>58.6</p></td><td><p>< 0.001</p></td><td><p>58.55</p></td><td char="." align="char"><p>6.82</p></td><td char="." align="char"><p>0.009</p></td></tr><tr><td><p> Speed</p></td><td><p>14</p></td><td><p>1.521</p></td><td><p>1.203–1.839</p></td><td><p>76.2</p></td><td><p>< 0.001</p></td><td><p>76.21</p></td><td /><td /></tr></tbody></table> </ephtml> </p> <p>Number of independent effect sizes (<emph>k</emph>) that contributed to each meta-analysis; the weighted average effect size of group differences (<emph>d</emph>); 95% confidence interval (CI); <emph>Z</emph> significance test statistic for the effect sizes and corresponding <emph>p</emph><emph> value</emph>; the proportion of total variance between effect sizes not explained by chance (<emph>I</emph><sups>2</sups>); between-groups heterogeneity of variance (<emph>Q</emph><subs>Between</subs>) and corresponding <emph>p value</emph></p> <p>Likewise, on reading- and writing-related variables, ESs for both phonological awareness (accuracy: <emph>d</emph> = 1.074; speed: <emph>d</emph> = 1.404) and orthographic knowledge (accuracy: <emph>d</emph> = 1.027; speed: <emph>d</emph> = 1.521) were larger when speed measures were used instead of accuracy (<emph>p</emph> < 0.001 and <emph>p</emph> = 0.009, respectively).</p> <p>For the remaining variables, either the number of available studies was insufficient, or the assessments were based on either speed or accuracy measures.</p> <p> <emph>To what extent does orthographic transparency modulate the reading profile of adults with dyslexia?</emph> </p> <p>To determine whether orthographic transparency modulates the way in which symptoms are expressed in adults with dyslexia, we separated this moderator variable into three categories of transparency (opaque, intermediate and transparent). As seen in Table 3, the three orthographies displayed significant ESs (most of them very large) across all reading and reading-related skills that we examined. These results confirm that, independent of the properties of the orthographies, dyslexia symptoms persist into adulthood. However, the magnitude of some of these deficits does vary.</p> <p>Effect sizes with 95% confidence intervals and heterogeneity between orthographies in studies comparing adults with dyslexia with typical readers</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th rowspan="2" /><th rowspan="2"><p><italic>k</italic></p></th><th colspan="2"><p>Effect size</p></th><th rowspan="2"><p><italic>Z</italic></p></th><th rowspan="2"><p><italic>p</italic> value</p></th><th colspan="3"><p>Heterogeneity</p></th></tr><tr><th><p><italic>d</italic></p></th><th><p>95% CI</p></th><th><p><italic>I</italic><sup>2</sup></p></th><th><p><italic>Q</italic><sub>Between</sub></p></th><th><p><italic>p</italic> value</p></th></tr></thead><tbody><tr><td colspan="9"><p>Reading and writing variables</p></td></tr><tr><td colspan="9"><p> Word read (acc)</p></td></tr><tr><td><p> Opaque</p></td><td><p>81</p></td><td><p>1.619 <sup>b</sup></p></td><td><p>1.452–1.787</p></td><td><p>18.6</p></td><td><p>< 0.001</p></td><td><p>77.97</p></td><td char="." align="char"><p>8.27</p></td><td char="." align="char"><p>0.016</p></td></tr><tr><td><p> Intermediate</p></td><td><p>11</p></td><td><p>1.420 <sup>ab</sup></p></td><td><p>1.010–1.830</p></td><td><p>6.8</p></td><td><p>< 0.001</p></td><td><p>72.67</p></td><td /><td /></tr><tr><td><p> Transparent</p></td><td><p>14</p></td><td><p>1.070 <sup>a</sup></p></td><td><p>0.730–1.409</p></td><td><p>6.2</p></td><td><p>< 0.001</p></td><td><p>76.08</p></td><td /><td /></tr><tr><td colspan="9"><p> Word read (speed)</p></td></tr><tr><td><p> Opaque</p></td><td><p>85</p></td><td><p>1.952 <sup>a</sup></p></td><td><p>1.775–2.130</p></td><td><p>21.6</p></td><td><p>< 0.001</p></td><td><p>79.52</p></td><td char="." align="char"><p>1.94</p></td><td char="." align="char"><p>0.378</p></td></tr><tr><td><p> Intermediate</p></td><td><p>29</p></td><td><p>1.938 <sup>a</sup></p></td><td><p>1.629–2.247</p></td><td><p>12.3</p></td><td><p>< 0.001</p></td><td><p>85.72</p></td><td /><td /></tr><tr><td><p> Transparent</p></td><td><p>16</p></td><td><p>1.672 <sup>a</sup></p></td><td><p>1.315–2.030</p></td><td><p>9.2</p></td><td><p>< 0.001</p></td><td><p>76.32</p></td><td /><td /></tr><tr><td colspan="9"><p> Pseudoword read (acc)</p></td></tr><tr><td><p> Opaque</p></td><td><p>62</p></td><td><p>1.738 <sup>b</sup></p></td><td><p>1.552–1.923</p></td><td><p>18.4</p></td><td><p>< 0.001</p></td><td><p>74.82</p></td><td char="." align="char"><p>5.23</p></td><td char="." align="char"><p>0.073</p></td></tr><tr><td><p> Intermediate</p></td><td><p>11</p></td><td><p>1.976 <sup>b</sup></p></td><td><p>1.373–2.579</p></td><td><p>6.4</p></td><td><p>< 0.001</p></td><td><p>85.86</p></td><td /><td /></tr><tr><td><p> Transparent</p></td><td><p>11</p></td><td><p>1.304 <sup>a</sup></p></td><td><p>0.933–1.675</p></td><td><p>6.9</p></td><td><p>< 0.001</p></td><td><p>72.67</p></td><td /><td /></tr><tr><td colspan="9"><p> Pseudoword read (speed)</p></td></tr><tr><td><p> Opaque</p></td><td><p>77</p></td><td><p>2.043 <sup>b</sup></p></td><td><p>1.869–2.216</p></td><td><p>23.0</p></td><td><p>< 0.001</p></td><td><p>76.70</p></td><td char="." align="char"><p>6.09</p></td><td char="." align="char"><p>0.048</p></td></tr><tr><td><p> Intermediate</p></td><td><p>27</p></td><td><p>2.354 <sup>b</sup></p></td><td><p>1.963–2.745</p></td><td><p>11.8</p></td><td><p>< 0.001</p></td><td><p>89.61</p></td><td /><td /></tr><tr><td><p> Transparent</p></td><td><p>13</p></td><td><p>1.711 <sup>a</sup></p></td><td><p>1.374–2.048</p></td><td><p>10.0</p></td><td><p>< 0.001</p></td><td><p>71.96</p></td><td /><td /></tr><tr><td colspan="9"><p> Reading comprehension (acc)</p></td></tr><tr><td><p> Opaque</p></td><td><p>20</p></td><td><p>0.614 <sup>a</sup></p></td><td><p>0.404–0.824</p></td><td><p>5.7</p></td><td><p>< 0.001</p></td><td><p>65.82</p></td><td char="." align="char"><p>6.13</p></td><td char="." align="char"><p>0.047</p></td></tr><tr><td><p> Intermediate</p></td><td><p>8</p></td><td><p>1.354 <sup>b</sup></p></td><td><p>0.735–1.973</p></td><td><p>4.3</p></td><td><p>< 0.001</p></td><td><p>84.16</p></td><td /><td /></tr><tr><td><p> Transparent</p></td><td><p>8</p></td><td><p>0.452 <sup>a</sup></p></td><td><p>0.074–0.829</p></td><td><p>2.4</p></td><td><p>0.019</p></td><td><p>79.40</p></td><td /><td /></tr><tr><td colspan="9"><p> Spelling (acc)</p></td></tr><tr><td><p> Opaque</p></td><td><p>60</p></td><td><p>1.708 <sup>b</sup></p></td><td><p>1.556–1.861</p></td><td><p>22.0</p></td><td><p>< 0.001</p></td><td><p>61.78</p></td><td char="." align="char"><p>9.97</p></td><td char="." align="char"><p>0.007</p></td></tr><tr><td><p> Intermediate</p></td><td><p>22</p></td><td><p>2.073 <sup>c</sup></p></td><td><p>1.716–2.430</p></td><td><p>11.4</p></td><td><p>< 0.001</p></td><td><p>85.46</p></td><td /><td /></tr><tr><td><p> Transparent</p></td><td><p>14</p></td><td><p>1.250 <sup>a</sup></p></td><td><p>0.882–1.618</p></td><td><p>6.7</p></td><td><p>< 0.001</p></td><td><p>84.14</p></td><td /><td /></tr><tr><td colspan="9"><p>Reading- and writing-related variables</p></td></tr><tr><td colspan="9"><p> Phonological awareness (acc)</p></td></tr><tr><td><p> Opaque</p></td><td><p>71</p></td><td><p>1.145 <sup>b</sup></p></td><td><p>1.019–1.271</p></td><td><p>17.9</p></td><td><p>< 0.001</p></td><td><p>60.77</p></td><td char="." align="char"><p>10.79</p></td><td char="." align="char"><p>0.005</p></td></tr><tr><td><p> Intermediate</p></td><td><p>15</p></td><td><p>1.264 <sup>b</sup></p></td><td><p>0.975–1.552</p></td><td><p>8.6</p></td><td><p>< 0.001</p></td><td><p>69.24</p></td><td /><td /></tr><tr><td><p> Transparent</p></td><td><p>10</p></td><td><p>0.633 <sup>a</sup></p></td><td><p>0.328–0.937</p></td><td><p>4.1</p></td><td><p>< 0.001</p></td><td><p>68.20</p></td><td /><td /></tr><tr><td colspan="9"><p> Phonological memory (acc)</p></td></tr><tr><td><p> Opaque</p></td><td><p>40</p></td><td><p>1.052 <sup>a</sup></p></td><td><p>0.891–1.212</p></td><td><p>12.9</p></td><td><p>< 0.001</p></td><td><p>57.22</p></td><td char="." align="char"><p>0.25</p></td><td char="." align="char"><p>0.884</p></td></tr><tr><td><p> Intermediate</p></td><td><p>18</p></td><td><p>1.048 <sup>a</sup></p></td><td><p>0.852–1.244</p></td><td><p>10.5</p></td><td><p>< 0.001</p></td><td><p>52.61</p></td><td /><td /></tr><tr><td><p> Transparent</p></td><td><p>8</p></td><td><p>0.944 <sup>a</sup></p></td><td><p>0.545–1.344</p></td><td><p>4.6</p></td><td><p>< 0.001</p></td><td><p>73.06</p></td><td /><td /></tr><tr><td colspan="9"><p> RAN (speed)</p></td></tr><tr><td><p> Opaque</p></td><td><p>54</p></td><td><p>1.211 <sup>a</sup></p></td><td><p>1.100–1.323</p></td><td><p>21.3</p></td><td><p>< 0.001</p></td><td><p>24.40</p></td><td char="." align="char"><p>0.26</p></td><td char="." align="char"><p>0.880</p></td></tr><tr><td><p> Intermediate</p></td><td><p>16</p></td><td><p>1.175 <sup>a</sup></p></td><td><p>0.980–1.370</p></td><td><p>11.8</p></td><td><p>< 0.001</p></td><td><p>54.14</p></td><td /><td /></tr><tr><td><p> Transparent</p></td><td><p>7</p></td><td><p>1.110 <sup>a</sup></p></td><td><p>0.655–1.565</p></td><td><p>4.8</p></td><td><p>< 0.001</p></td><td><p>77.91</p></td><td /><td /></tr></tbody></table> </ephtml> </p> <p>Number of independent effect sizes (<emph>k</emph>) that contributed to each meta-analysis; the weighted average effect size of group differences (<emph>d</emph>); 95% confidence interval (CI); <emph>Z</emph> significance test statistic for the effect sizes and corresponding <emph>p value</emph>; the proportion of total variance between effect sizes not explained by chance (<emph>I</emph><sups>2</sups>); between-groups heterogeneity of variance (<emph>Q</emph><subs>Between</subs>) and corresponding <emph>p value</emph>; mean effect sizes not sharing the same superscript letters (<sups>a–c</sups>) differ at <emph>p</emph> < 0.05</p> <p>When the opacity of the orthography was considered (see Table 3), significant differences were observed for word reading accuracy (<emph>p</emph> = 0.016). Transparent had the lowest ES (<emph>d</emph> = 1.070), followed by intermediate (<emph>d</emph> = 1.420) and opaque orthography (<emph>d</emph> = 1.619). While the difference between transparent and opaque was significant (<emph>p</emph> = 0.004), no differences between transparent and intermediate and between intermediate and opaque were observed. When word reading was measured with speed, all orthographies revealed larger ESs (transparent: <emph>d</emph> = 1.672; intermediate: <emph>d</emph> = 1.938; opaque: <emph>d</emph> = 1.952), though no significant differences between the three levels of orthographic complexity were registered.</p> <p>For pseudoword reading accuracy, orthographic transparency yielded a marginally significant effect (<emph>p</emph> = 0.073). Again, transparent orthographies showed the smallest ES (<emph>d</emph> = 1.304), followed by the opaque ES (<emph>d</emph> = 1.738) and the intermediate orthographies ES (<emph>d</emph> = 1.976). Pairwise comparisons showed that the difference between transparent vs. opaque (<emph>p</emph> = 0.041) and between transparent vs. intermediate (<emph>p</emph> = 0.063) orthographies approached significance. Larger ESs were observed for pseudoword speed compared with pseudoword accuracy (transparent: <emph>d</emph> = 1.711; opaque: <emph>d</emph> = 2.043; intermediate: <emph>d</emph> = 2.354) with significant differences among the three orthographies (<emph>p</emph> = 0.048). Pairwise comparisons showed a difference between transparent vs. intermediate (<emph>p</emph> = 0.015) and transparent vs. opaque marginally significant (<emph>p</emph> = 0.087) orthographies with opaque and intermediate performing at an equivalent level.</p> <p>Reading comprehension accuracy registered significant differences among orthographies (<emph>p</emph> = 0.047) with larger ESs for intermediate orthographies (<emph>d</emph> = 1.354) compared with transparent (<emph>d</emph> = 0.452) and opaque (<emph>d</emph> = 0.614). Pairwise comparisons showed that the deficit in reading comprehension was significantly more marked for intermediate orthographies (in contrast to transparent orthographies, <emph>p</emph> = 0.015, and in contrast to opaque orthographies, <emph>p</emph> = 0.066).</p> <p>The same holds true for spelling accuracy (<emph>p</emph> = 0.007) with larger ESs in studies sampling participants from intermediate orthographies (intermediate: <emph>d</emph> = 2.073; opaque: <emph>d</emph> = 1.708; transparent: <emph>d</emph> = 1.250). Pairwise contrasts between transparent vs. intermediate (<emph>p</emph> < 0.002), transparent vs. opaque (<emph>p</emph> = 0.024) and intermediate vs. opaque (<emph>p</emph> = 0.066) were all reliable.</p> <p>Concerning reading- and writing-related skills, orthographic consistency is a significant moderator variable, in particular for phonological awareness (<emph>p</emph> = 0.005) with a medium-to-large ES observed in transparent orthographies (<emph>d</emph> = 0.633) and rather large ESs observed in intermediate and opaque orthographies (<emph>d</emph> = 1.264 and <emph>d</emph> = 1.145, respectively). Pairwise comparisons confirmed that the deficits in this measure were smaller for adults with dyslexia from transparent orthographies (in contrast to intermediate orthographies, <emph>p</emph> = 0.003, and in contrast to opaque orthographies, <emph>p</emph> = 0.002, while no differences were observed for intermediate vs. opaque orthographies).</p> <p>In terms of phonological memory, no significant moderator effect of orthography was observed. The three orthographic groups showed rather equivalent ESs (transparent: <emph>d</emph> = 0.944; intermediate: <emph>d</emph> = 1.048; opaque: <emph>d</emph> = 1.052). The same was true for RAN (transparent: <emph>d</emph> = 1.110; intermediate: <emph>d</emph> = 1.175; opaque: <emph>d</emph> = 1.211).</p> <p>For the remaining variables, the impact of orthography on adults with dyslexia could not be evaluated using the present data set, given the reduced number of effect sizes available per category.</p> <hd id="AN0146733574-31">Regression analysis</hd> <p>Lastly, given that age may be associated with a practice factor, a meta-regression analysis was performed to examine whether this variable predicts the size of the observed deficits in individuals with dyslexia. This analysis included only studies that mentioned a mean age value for the dyslexic sample. It is apparent in Table 4, that even in adulthood, age predicts the size of the reading deficit in dyslexia but only in terms of the accuracy measures of real word reading (<emph>β</emph> = − 0.04; <emph>p</emph> = 0.01). This is also true for spelling (<emph>β</emph> = − 0.03; <emph>p</emph> = 0.03), while age has no significant impact on the ES obtained for the other measures assessed.</p> <p>Regression analysis for age as moderator in studies comparing adults with dyslexia with typical readers</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th><p>Continuous moderators</p></th><th><p><italic>k</italic></p></th><th><p>Slope <italic>β</italic></p></th><th><p><italic>p</italic> value</p></th><th><p><italic>R</italic><sup>2</sup></p></th></tr></thead><tbody><tr><td colspan="5"><p>Age (dyslexic group)</p></td></tr><tr><td colspan="5"><p> Word read</p></td></tr><tr><td><p> Accuracy</p></td><td><p>103</p></td><td><p>− 0.04</p></td><td><p>0.01</p></td><td><p>0.05</p></td></tr><tr><td><p> Speed</p></td><td><p>121</p></td><td><p>− 0.02</p></td><td><p>0.31</p></td><td><p>0.00</p></td></tr><tr><td colspan="5"><p> Pseudoword read</p></td></tr><tr><td><p> Accuracy</p></td><td><p>82</p></td><td><p>− 0.03</p></td><td><p>0.08</p></td><td><p>0.00</p></td></tr><tr><td><p> Speed</p></td><td><p>109</p></td><td><p>− 0.00</p></td><td><p>0.82</p></td><td><p>0.00</p></td></tr><tr><td colspan="5"><p> Text read</p></td></tr><tr><td><p> Accuracy</p></td><td><p>20</p></td><td><p>− 0.06</p></td><td><p>0.11</p></td><td><p>0.00</p></td></tr><tr><td><p> Speed</p></td><td><p>35</p></td><td><p>− 0.04</p></td><td><p>0.19</p></td><td><p>0.00</p></td></tr><tr><td colspan="5"><p> Reading comprehension</p></td></tr><tr><td><p> Accuracy</p></td><td><p>34</p></td><td><p>− 0.02</p></td><td><p>0.16</p></td><td><p>0.02</p></td></tr><tr><td colspan="5"><p> Spelling</p></td></tr><tr><td><p> Accuracy</p></td><td><p>92</p></td><td><p>− 0.03</p></td><td><p>0.03</p></td><td><p>0.07</p></td></tr><tr><td colspan="5"><p> Phonological awareness</p></td></tr><tr><td><p> Accuracy</p></td><td><p>93</p></td><td><p>− 0.00</p></td><td><p>0.77</p></td><td><p>0.00</p></td></tr><tr><td><p> Speed</p></td><td><p>46</p></td><td><p>0.01</p></td><td><p>0.57</p></td><td><p>0.00</p></td></tr><tr><td colspan="5"><p> Orthographic knowledge</p></td></tr><tr><td><p> Accuracy</p></td><td><p>20</p></td><td><p>− 0.02</p></td><td><p>0.37</p></td><td><p>0.00</p></td></tr><tr><td><p> Speed</p></td><td><p>13</p></td><td><p>− 0.03</p></td><td><p>0.22</p></td><td><p>0.02</p></td></tr><tr><td><p> RAN alphabetic (speed)</p></td><td><p>50</p></td><td><p>− 0.03</p></td><td><p>0.13</p></td><td><p>0.00</p></td></tr><tr><td><p> RAN non-alphabetic (speed)</p></td><td><p>39</p></td><td><p>0.01</p></td><td><p>0.78</p></td><td><p>0.00</p></td></tr></tbody></table> </ephtml> </p> <p>Number of independent effect sizes (<emph>k</emph>) that contributed to each meta-regression analysis; slope <emph>β</emph> and its corresponding <emph>p value</emph>; explanatory value of the moderator for differences between studies (<emph>R</emph><sups>2</sups>)</p> <hd id="AN0146733574-32">Sample characteristics</hd> <p>Considering the large pool of papers included in this meta-analysis and consequently the diversity of methods for identifying reading disabilities as well as other comorbidities, it would make sense to also examine if the different sample selection criteria might affect the results. As reported previously, 84.8% of the studies explicitly stated that participants included in the dyslexic group had a previous formal diagnosis, while the remaining did not use that criterion for including participants in the dyslexic group. Another possible source of variability is the way that authors dealt with comorbidities, namely the attentional disorders (about 58% of the studies explicitly stated that participants with attention problems were excluded). Therefore, the data was analysed in order to test if these different inclusion criteria do affect the ES of the main reading and writing variables.</p> <p>First, we examined if the presence of a previous diagnosis of dyslexia (151 studies and 156 samples) vs. no previous diagnosis (self-reported dyslexia or dyslexic participants identified and selected from a large pool of participants; 27 studies and 29 samples) had an impact on reading and writing measures. As can be seen in Table 5, only the deficit on pseudoword reading speed was moderated by the type of diagnosis; this deficit was larger in studies where dyslexic participants did not have a previous formal diagnosis. To determine the effect of the control of attentional deficits, we compared studies which explicitly claimed the exclusion of participants with attentional problems (103 studies and 105 samples) vs. studies that did not mention whether their samples included participants with attentional problems (72 studies and 77 samples). The results showed that controlling or not controlling the presence of attentional deficits does not moderate the ES in any of the reading and writing measures.</p> <p>Effect sizes with 95% confidence intervals and heterogeneity between subgroups (with vs. without previous diagnosis of dyslexia; with vs. without control of attentional deficits) in studies comparing adults with dyslexia with typical readers</p> <p> <ephtml> <table frame="hsides" rules="groups"><thead><tr><th rowspan="2" /><th rowspan="2"><p><italic>k</italic></p></th><th colspan="2"><p>Effect size</p></th><th rowspan="2"><p><italic>Z</italic></p></th><th rowspan="2"><p><italic>p</italic> value</p></th><th colspan="3"><p>Heterogeneity</p></th></tr><tr><th><p><italic>d</italic></p></th><th><p>95% CI</p></th><th><p><italic>I</italic><sup>2</sup></p></th><th><p><italic>Q</italic><sub>Between</sub></p></th><th><p><italic>p</italic> value</p></th></tr></thead><tbody><tr><td colspan="9"><p>Previous diagnosis of dyslexia</p></td></tr><tr><td colspan="9"><p> Word read (acc)</p></td></tr><tr><td><p> Yes</p></td><td><p>88</p></td><td><p>1.508</p></td><td><p>1.354–1.663</p></td><td><p>19.1</p></td><td><p>< 0.001</p></td><td><p>76.72</p></td><td char="." align="char"><p>0.12</p></td><td char="." align="char"><p>0.730</p></td></tr><tr><td><p> No</p></td><td><p>18</p></td><td><p>1.582</p></td><td><p>1.192–1.973</p></td><td><p>7.9</p></td><td><p>< 0.001</p></td><td><p>84.54</p></td><td /><td /></tr><tr><td colspan="9"><p> Word read (speed)</p></td></tr><tr><td><p> Yes</p></td><td><p>108</p></td><td><p>1.880</p></td><td><p>1.723–2.037</p></td><td><p>23.5</p></td><td><p>< 0.001</p></td><td><p>81.04</p></td><td char="." align="char"><p>1.21</p></td><td char="." align="char"><p>0.271</p></td></tr><tr><td><p> No</p></td><td><p>22</p></td><td><p>2.080</p></td><td><p>1.760–2.399</p></td><td><p>12.8</p></td><td><p>< 0.001</p></td><td><p>79.60</p></td><td /><td /></tr><tr><td colspan="9"><p> Pseudoword read (acc)</p></td></tr><tr><td><p> Yes</p></td><td><p>71</p></td><td><p>1.641</p></td><td><p>1.472–1.810</p></td><td><p>19.0</p></td><td><p>< 0.001</p></td><td><p>74.40</p></td><td char="." align="char"><p>1.87</p></td><td char="." align="char"><p>0.172</p></td></tr><tr><td><p> No</p></td><td><p>13</p></td><td><p>2.001</p></td><td><p>1.513–2.489</p></td><td><p>8.0</p></td><td><p>< 0.001</p></td><td><p>84.31</p></td><td /><td /></tr><tr><td colspan="9"><p> Pseudoword read (speed)</p></td></tr><tr><td><p> Yes</p></td><td><p>100</p></td><td><p>1.993</p></td><td><p>1.845–2.140</p></td><td><p>26.4</p></td><td><p>< 0.001</p></td><td><p>77.44</p></td><td char="." align="char"><p>3.91</p></td><td char="." align="char"><p>0.048</p></td></tr><tr><td><p> No</p></td><td><p>17</p></td><td><p>2.568</p></td><td><p>2.017–3.119</p></td><td><p>9.1</p></td><td><p>< 0.001</p></td><td><p>90.42</p></td><td /><td /></tr><tr><td colspan="9"><p> Text read (speed)</p></td></tr><tr><td><p> Yes</p></td><td><p>28</p></td><td><p>1.727</p></td><td><p>1.399–2.056</p></td><td><p>10.3</p></td><td><p>< 0.001</p></td><td><p>86.07</p></td><td char="." align="char"><p>0.06</p></td><td char="." align="char"><p>0.806</p></td></tr><tr><td><p> No</p></td><td><p>7</p></td><td><p>1.863</p></td><td><p>0.831–2.895</p></td><td><p>3.5</p></td><td><p>< 0.001</p></td><td><p>93.14</p></td><td /><td /></tr><tr><td colspan="9"><p> Reading comprehension (acc)</p></td></tr><tr><td><p> Yes</p></td><td><p>27</p></td><td><p>0.633</p></td><td><p>0.427–0.839</p></td><td><p>6.0</p></td><td><p>< 0.001</p></td><td><p>73.36</p></td><td char="." align="char"><p>1.76</p></td><td char="." align="char"><p>0.184</p></td></tr><tr><td><p> No</p></td><td><p>9</p></td><td><p>0.915</p></td><td><p>0.553–1.277</p></td><td><p>5.0</p></td><td><p>< 0.001</p></td><td><p>73.43</p></td><td /><td /></tr><tr><td colspan="9"><p> Spelling (acc)</p></td></tr><tr><td><p> Yes</p></td><td><p>80</p></td><td><p>1.752</p></td><td><p>1.582–1.923</p></td><td><p>20.1</p></td><td><p>< 0.001</p></td><td><p>79.89</p></td><td char="." align="char"><p>1.07</p></td><td char="." align="char"><p>0.301</p></td></tr><tr><td><p> No</p></td><td><p>16</p></td><td><p>1.581</p></td><td><p>1.305–1.857</p></td><td><p>11.2</p></td><td><p>< 0.001</p></td><td><p>69.11</p></td><td /><td /></tr><tr><td colspan="9"><p>Control of attentional deficits</p></td></tr><tr><td colspan="9"><p> Word read (acc)</p></td></tr><tr><td><p> Yes</p></td><td><p>60</p></td><td><p>1.436</p></td><td><p>1.259–1.614</p></td><td><p>15.9</p></td><td><p>< 0.001</p></td><td><p>73.87</p></td><td char="." align="char"><p>1.04</p></td><td char="." align="char"><p>0.309</p></td></tr><tr><td><p> No</p></td><td><p>45</p></td><td><p>1.587</p></td><td><p>1.357–1.818</p></td><td><p>13.5</p></td><td><p>< 0.001</p></td><td><p>81.16</p></td><td /><td /></tr><tr><td colspan="9"><p> Word read (speed)</p></td></tr><tr><td><p> Yes</p></td><td><p>78</p></td><td><p>1.868</p></td><td><p>1.691–2.045</p></td><td><p>20.7</p></td><td><p>< 0.001</p></td><td><p>78.63</p></td><td char="." align="char"><p>0.62</p></td><td char="." align="char"><p>0.432</p></td></tr><tr><td><p> No</p></td><td><p>50</p></td><td><p>1.989</p></td><td><p>1.743–2.236</p></td><td><p>15.8</p></td><td><p>< 0.001</p></td><td><p>84.43</p></td><td /><td /></tr><tr><td colspan="9"><p> Pseudoword read (acc)</p></td></tr><tr><td><p> Yes</p></td><td><p>54</p></td><td><p>1.669</p></td><td><p>1.478–1.860</p></td><td><p>17.1</p></td><td><p>< 0.001</p></td><td><p>73.79</p></td><td char="." align="char"><p>0.29</p></td><td char="." align="char"><p>0.590</p></td></tr><tr><td><p> No</p></td><td><p>30</p></td><td><p>1.768</p></td><td><p>1.460–2.077</p></td><td><p>11.2</p></td><td><p>< 0.001</p></td><td><p>82.21</p></td><td /><td /></tr><tr><td colspan="9"><p> Pseudoword read (speed)</p></td></tr><tr><td><p> Yes</p></td><td><p>72</p></td><td><p>2.016</p></td><td><p>1.840–2.192</p></td><td><p>22.5</p></td><td><p>< 0.001</p></td><td><p>77.89</p></td><td char="." align="char"><p>0.91</p></td><td char="." align="char"><p>0.339</p></td></tr><tr><td><p> No</p></td><td><p>43</p></td><td><p>2.182</p></td><td><p>1.891–2.473</p></td><td><p>14.7</p></td><td><p>< 0.001</p></td><td><p>86.16</p></td><td /><td /></tr><tr><td colspan="9"><p> Text read (speed)</p></td></tr><tr><td><p> Yes</p></td><td><p>18</p></td><td><p>1.831</p></td><td><p>1.362–2.299</p></td><td><p>7.7</p></td><td><p>< 0.001</p></td><td><p>87.34</p></td><td char="." align="char"><p>0.13</p></td><td char="." align="char"><p>0.719</p></td></tr><tr><td><p> No</p></td><td><p>17</p></td><td><p>1.710</p></td><td><p>1.250–2.171</p></td><td><p>7.3</p></td><td><p>< 0.001</p></td><td><p>89.34</p></td><td /><td /></tr><tr><td colspan="9"><p> Reading comprehension (acc)</p></td></tr><tr><td><p> Yes</p></td><td><p>20</p></td><td><p>0.600</p></td><td><p>0.342–0.859</p></td><td><p>4.5</p></td><td><p>< 0.001</p></td><td><p>72.29</p></td><td char="." align="char"><p>1.41</p></td><td char="." align="char"><p>0.235</p></td></tr><tr><td><p> No</p></td><td><p>16</p></td><td><p>0.822</p></td><td><p>0.563–1.082</p></td><td><p>6.2</p></td><td><p>< 0.001</p></td><td><p>77.12</p></td><td /><td /></tr><tr><td colspan="9"><p> Spelling (acc)</p></td></tr><tr><td><p> Yes</p></td><td><p>48</p></td><td><p>1.705</p></td><td><p>1.471–1.940</p></td><td><p>14.3</p></td><td><p>< 0.001</p></td><td><p>82.29</p></td><td char="." align="char"><p>0.02</p></td><td char="." align="char"><p>0.879</p></td></tr><tr><td><p> No</p></td><td><p>45</p></td><td><p>1.729</p></td><td><p>1.543–1.914</p></td><td><p>18.2</p></td><td><p>< 0.001</p></td><td><p>71.99</p></td><td /><td /></tr></tbody></table> </ephtml> </p> <p>Number of independent effect sizes (<emph>k</emph>) that contributed to each meta-analysis; the weighted average effect size of group differences (<emph>d</emph>); 95% confidence interval (CI); <emph>Z</emph> significance test statistic for the effect sizes and corresponding <emph>p value</emph>; the proportion of total variance between effect sizes not explained by chance (<emph>I</emph><sups>2</sups>); between-groups heterogeneity of variance (<emph>Q</emph><subs>Between</subs>) and corresponding <emph>p value</emph></p> <hd id="AN0146733574-33">Discussion</hd> <p>Dyslexia is a disorder that persists until adulthood, yet studies on adults with dyslexia are scarce as compared with studies on children with dyslexia. In this meta-analysis, we aimed to perform a systematic characterisation of the weaknesses that are primarily associated with dyslexia in adulthood. The knowledge available in the literature suggests that problems in reading fluency and spelling are the most prominent markers of dyslexia in adulthood (Nergård-Nilssen and Hulme [<reflink idref="bib37" id="ref65">37</reflink>]). On the other hand, some studies indicate that adults with dyslexia still differ significantly from those without dyslexia in terms of several cognitive skills, such as phonological short-term memory, phonological awareness and whole-word processing (Tamboer et al., [<reflink idref="bib58" id="ref66">58</reflink>]; Swanson & Hsieh, [<reflink idref="bib56" id="ref67">56</reflink>]). However, with few exceptions, most of these findings come from studies conducted in English, an opaque orthography. This is critical, because it is well-established that orthographic transparency has a strong impact on the rate at which reading skills develop and the way in which symptoms of reading difficulties are expressed (Landerl et al., [<reflink idref="bib30" id="ref68">30</reflink>]; Share, [<reflink idref="bib53" id="ref69">53</reflink>]). Furthermore, the assessment of adults with dyslexia has used either accuracy or speed, or both, as performance indices. However, it is questionable whether these indices have the same sensitivity in detecting reading disorders in adulthood and if they potentially interact with orthographic transparency. To address these issues and to obtain a characterisation of the major cognitive symptoms of dyslexia in adults, we conducted a meta-analytic review that added performance indices (accuracy and speed) and orthographic transparency as moderator variables.</p> <hd id="AN0146733574-34">How do adults with dyslexia differ from adult typical readers?</hd> <p>Our first question sought to analyse the extent to which adults with dyslexia differ from typical readers (control group) on measures of reading and writing (word, pseudoword, and text reading, reading comprehension, and spelling) and related cognitive skills (phonological awareness, phonological memory, verbal working memory, orthographic knowledge, vocabulary and rapid automatised naming) as well as on general cognitive abilities (verbal and non-verbal IQs, abstraction and speed of processing). Empirical studies focusing on adult groups of readers with dyslexia most often assess the performance of high-functioning, university students who were diagnosed in childhood, many of whom had probably participated in intervention programmes. In this review, 80% of the studies tested university students, so we could expect that some cognitive domains would be compensated for to varying degrees and, consequently, the differences compared with typical readers attenuated.</p> <p>Overall, our meta-analytic results showed that adults with dyslexia still score more poorly than typical readers on reading measures, even when context can assist in decoding as in text reading, as well as on all reading-related measures considered in this study. Importantly, this was generally expressed by large effect sizes across domains (most of them above 1.0) with a few exceptions (vocabulary and reading comprehension). Hence, many cognitive–linguistic constructs are implicated in dyslexia and impairments in the cognitive processes that underlie reading skills persist into adulthood. Interestingly, symptoms are relatively more severe for reading and writing abilities (median <emph>d</emph> = 1.761; ranging from <emph>d</emph> = 0.729 to <emph>d</emph> = 2.034) than for the cognitive processing skills associated with literacy (median <emph>d</emph> = 1.103; ranging from <emph>d</emph> = 0.591 to <emph>d</emph> = 1.295), consistent with Swanson and Hsieh's ([<reflink idref="bib56" id="ref70">56</reflink>]) prior meta-analysis. This dissociation may suggest that years of reading exposure, together with remediation programmes, could help improve cognitive processing abilities that support reading and writing (although not sufficiently to perform at a normal level). An alternative explanation could be that, in adulthood, reading and writing are less dependent on these abilities than they are in children.</p> <p>The only non-significant comparison between typical readers and the dyslexic group was for abstraction, whereas small and small-to-medium ESs were registered for measures related to general cognitive skills, such as non-verbal IQ (<emph>d</emph> = 0.187), full IQ (<emph>d</emph> = 0.296) and verbal IQ (<emph>d</emph> = 0.424). In our results, the relatively larger difference registered for verbal IQ is probably due to the fact that dyslexic individuals have less experience with reading and writing and, therefore, have a lag in verbal ability. Still, their full IQ score was within the population average (dyslexic mean = 110.9, ranging from 100.5 to 126.0; controls mean = 113.5, ranging from 104.9 to 128.8). A practical implication of these results is that non-verbal IQ measures should be preferable when assessing general cognitive function in individuals with dyslexia, as these measures depend less on reading abilities and therefore are less detrimental to these readers. It is important to note, however, that the reported differences between groups in general cognitive functioning, as measured by non-verbal IQ, did not bias the estimated word reading deficit (meta-regression for word reading accuracy: <emph>k</emph> = 72, <emph>β</emph> = − 0.121, <emph>p</emph> = 0.642, <emph>R</emph><sups>2</sups> = 0.00; for word reading speed: <emph>k</emph> = 82, <emph>β</emph> = − 0.128, <emph>p</emph> = 0.645, <emph>R</emph><sups>2</sups> = 0.00). This finding supports the uncoupling of reading and IQ (Ferrer, Shaywitz, Holahan, Marchione, & Shaywitz, [<reflink idref="bib18" id="ref71">18</reflink>]).</p> <p>We also investigated the extent to which the age of the dyslexic group (range 18 to 41 years) was a possible moderator of the effect sizes obtained. Significant negative effects of age were observed only on word reading accuracy and spelling accuracy, suggesting that the older the participants, the smaller the deficit. These effects are most likely explained by the reading practice associated with age. Interestingly, word reading and spelling correspond to those variables indicating that they may be more sensitive to lexical knowledge. This might suggest that the compensation observed in older participants results from the increase in their lexical knowledge.</p> <hd id="AN0146733574-35">To what extent do speed measures pose a greater challenge than accuracy measures in an adult...</hd> <p>An innovative aspect of this meta-analytic review was our intention to clarify whether and to what extent speed measures pose an added challenge, as compared with accuracy measures, to adults with dyslexia with years of print exposure. Some studies have suggested that while accuracy improves, adult dyslexics may still experience problems with reading fluency (Eloranta et al., [<reflink idref="bib16" id="ref72">16</reflink>]). Thus, for each domain, we compared the magnitude of the deficit in accuracy measures with the magnitude of the deficit in speed measures (e.g. word reading accuracy vs. word reading speed). Our results showed that, relative to control readers, dyslexic adults, indeed, perform more poorly on speed measures of word and pseudoword reading, phonological awareness and orthographic knowledge than on corresponding accuracy measures. The stronger ESs for speed measures indicate that automaticity is significantly associated with reading ability, but this is harder to attain in adults with dyslexia who, compared with children, have more years of reading experience. This result is relevant for neuropsychological/psychoeducational assessment. Given that their performance is less impaired when accuracy is measured, tests measuring speed are more sensitive to differences between adults with dyslexia and control samples. For text reading, reading comprehension and spelling, the magnitude of the deficit was, in turn, not significantly different between accuracy and speed measures.</p> <p>In sum, in adulthood, the main symptoms associated with dyslexia seem somehow attenuated (word and pseudoword reading, phonological awareness and orthographic knowledge) when performance is quantified using accuracy, and seem amplified when performance is measured with time. However, this dissociation between accuracy and speed may be inflated particularly in transparent orthographies, as we discuss below.</p> <hd id="AN0146733574-36">Does orthographic transparency modulate the reading profile of adults with dyslexia?</hd> <p>The third question we intend to examine in this meta-analytic review is the extent to which the orthography in which we learn to read and write modulates the reading profile of adults with dyslexia. During the last decade, a number of cross-cultural studies focused on the influence of orthographic consistency on reading development (Seymour et al., [<reflink idref="bib52" id="ref73">52</reflink>]), on reading predictors (e.g. Caravolas, Lervåg, Defior, Seidlová Málková, & Hulme, [<reflink idref="bib9" id="ref74">9</reflink>]; Moll et al., [<reflink idref="bib36" id="ref75">36</reflink>]; Vaessen et al., [<reflink idref="bib62" id="ref76">62</reflink>]; Ziegler et al., [<reflink idref="bib67" id="ref77">67</reflink>]) and on the expression of dyslexia (Becker et al., [<reflink idref="bib3" id="ref78">3</reflink>]; Landerl et al., [<reflink idref="bib30" id="ref79">30</reflink>]). Given this previous research, we expected that dyslexia's main deficits would be less prominent in orthographies where access to the written code is more straightforward. In more transparent orthographies, learning a limited number of grapheme–phoneme correspondences facilitates the reading of a great number of words, even without previous contact with the word, while in opaque orthographies, the individuals must learn and store a larger number of grapheme–phoneme units so that they can read and write effectively.</p> <p>Our results clearly showed that orthographic transparency is an important factor affecting the way deficits express themselves. In adulthood, dyslexic participants who learned how to read and write in transparent orthographies showed less marked deficits overall as compared with participants who learned to read in more opaque orthographies. This was particularly visible for accuracy reading measures that seem to discriminate against participants from different orthographies. Namely, we found significant differences for word and pseudoword reading, reading comprehension and spelling, which favoured transparent orthographies. On the other hand, deficits in speed measures were particularly high and relatively homogeneous, confirming that fluency is a major problem in adulthood dyslexia across orthographies. In general, these findings corroborate previous cross-language studies which have indicated that inaccurate decoding and slow reading characterise the manifestations of dyslexia in opaque orthographies, while slow and effortful reading, rather than poor accuracy, is characteristic of dyslexia in more transparent ones (e.g. Goulandris, [<reflink idref="bib22" id="ref80">22</reflink>]).</p> <p>For word reading accuracy, participants with dyslexia from transparent orthographies differ less from normative readers compared with those from opaque orthographies (although both ESs were large). The comparison between these two orthographic groups was also significant. These results are consistent with the literature suggesting that the more opaque the orthography, the greater the reading deficit (Paulesu et al., [<reflink idref="bib43" id="ref81">43</reflink>]; Ziegler et al., [<reflink idref="bib68" id="ref82">68</reflink>]). However, for word reading speed, we found no significant differences in the magnitude of the deficit between orthographic groups, suggesting that the reading fluency deficit in dyslexia is equivalent across orthographies. It is interesting to note that dyslexic readers from transparent orthographies are less impaired in terms of word reading accuracy (<emph>d</emph> = 1.070) as compared with word reading speed (<emph>d</emph> = 1.672). This suggests that speed measures are more suited than accuracy to identify reading problems in adults, particularly those coming from more transparent orthographies (Re et al., [<reflink idref="bib50" id="ref83">50</reflink>]; Suárez-Coalla & Cuetos [<reflink idref="bib15" id="ref84">15</reflink>]).</p> <p>The results from pseudoword reading followed a similar pattern to word reading for both accuracy and speed. When pseudoword accuracy was considered, the transparent orthography showed the lowest ES, followed by the opaque and the intermediate orthographies, with a slightly significant difference between these two systems. Again, learning to read in a transparent orthography seems to facilitate decoding. Just like word reading speed, pseudoword reading speed presented larger ESs as compared with accuracy and seems particularly difficult for dyslexic adults in intermediate and opaque orthographies (<emph>d</emph> > 2.0). In transparent orthographies, the pseudoword reading speed deficit was close to the word reading speed deficit (<emph>d</emph> ~ 1.7). Overall, it seems that while adults with dyslexia from more opaque orthographies still struggle with reading accuracy (word and pseudoword), all adults with dyslexia, independent of the transparency of the orthographic code, have reading speed problems that are more emphasised in decoding tasks. This outcome confirms previous studies performed in various languages [e.g. Finish (Lyytinen, Aro, & Holopainnen, [<reflink idref="bib33" id="ref85">33</reflink>]), Italian (Zoccolotti et al., [<reflink idref="bib69" id="ref86">69</reflink>]), Dutch (Yap & van der Leij, [<reflink idref="bib66" id="ref87">66</reflink>]) and Hebrew (Breznitz, [<reflink idref="bib7" id="ref88">7</reflink>])].</p> <p>Our study also confirmed that beyond reading difficulties, spelling represents a core problem in adults with dyslexia, whose magnitude is moderated by orthography. Participants with dyslexia from transparent orthographies have significantly less pronounced deficits in spelling compared with participants from the other orthographies. Data from several studies conducted in orthographies with different feedforward and feedback consistencies have previously suggested that spelling problems are the most prominent marker of dyslexia in adults (Callens et al., [<reflink idref="bib8" id="ref89">8</reflink>]; Everatt, [<reflink idref="bib17" id="ref90">17</reflink>]; Kemp et al., [<reflink idref="bib28" id="ref91">28</reflink>]; Nergård-Nilssen and Hulme [<reflink idref="bib37" id="ref92">37</reflink>]) and are highly discriminative for dyslexia in higher education (Lindgrén & Laine, [<reflink idref="bib31" id="ref93">31</reflink>]). This is at least partly expected because phoneme-to-grapheme mappings needed for writing are typically less predictable than grapheme-to-phoneme mappings needed for reading (Bosman & Van Orden, [<reflink idref="bib6" id="ref94">6</reflink>]). Consequently, accurate spelling is potentially more demanding than reading. Our results considering orthography as a moderator seem to support this view: spelling deficits were significantly smaller in dyslexic readers from transparent orthographies (i.e. with the highest cross-code consistency both in feedforward and feedback directions; <emph>d</emph> = 1.250) as compared with intermediate (<emph>d</emph> = 2.073) and opaque orthographies (<emph>d</emph> = 1.708).</p> <p>Indeed, adults with dyslexia (particularly high-functioning dyslexics) seem to be able to use phonological skills to spell words, but they have difficulty in memorising orthographic patterns (Kemp et al., [<reflink idref="bib28" id="ref95">28</reflink>]). While the use of these phonological strategies is particularly advantageous for transparent orthographies, the difficulty in memorising orthographic patterns prevents the accurate spelling of irregular words that depend on orthographic knowledge. Interestingly, our dyslexic sample revealed a deficit in the orthographic knowledge (<emph>d</emph> = 1.233). The existence of an orthographic deficit in our sample, together with the possibility of adopting a phonological strategy when spelling consistent words, can thus explain the smaller deficit observed in transparent compared with more opaque orthographies, where efficient spelling demands a more lexical strategy.</p> <p>The observed largest ES for spelling in intermediate orthographies might rely on the discrepancy between reading and writing in these orthographies. That is, while in opaque orthographies the inconsistent grapheme-to-phoneme correspondences somehow prepare the individuals to deal with the equivalent level of inconsistency involved in writing (namely, favouring lexical orthographic knowledge), in intermediate ones, the relatively consistent grapheme-to-phoneme correspondences do not favour compensatory strategies to deal with less transparent spelling. This may lead, in turn, to a larger spelling deficit, as we found.</p> <p>Reading comprehension seems to be the core reading skill for which the deficit in dyslexia is overall less marked (<emph>d</emph> = 0.729). For reading comprehension accuracy, we found a significant influence of orthographic transparency with deficit in adults with dyslexia being higher in intermediate compared with transparent and opaque orthographies. The effect of this moderator in reading comprehension has received little attention, especially in adults. To our knowledge, only two studies have analysed this question. Hanley, Masterson, Spencer and Evans ([<reflink idref="bib23" id="ref96">23</reflink>]), compared English to Welsh 5th grade children and found that the English children performed significantly better at answering comprehension questions about stories that they had read than the Welsh children. This result suggests that being a reader of a transparent orthography (Welsh) does not confer any advantage as far as reading comprehension is concerned, while the opaque orthography (English) seems to benefit comprehension. As the authors argued, the fact that the comprehension score of the English children actually exceeded the score of the Welsh children might conceivably occur because an opaque orthography fosters greater emphasis on semantics than on phonologically based reading strategies.</p> <p>More recently, Mcclung and Pearson ([<reflink idref="bib34" id="ref97">34</reflink>]) analysed the reading comprehension results from two international assessments (PIRLS and PISA) with children and adolescents from seven countries. The authors found that competent readers from opaque as compared with more transparent orthographies may experience smaller but beneficial effects of reading comprehension because opaque orthographies encourage a greater reliance on print-to-meaning connections (Perfetti & Stafura, [<reflink idref="bib46" id="ref98">46</reflink>]). On the contrary, reading comprehension of less competent readers may be slowed down by the demands of opaque orthographies (irregular phoneme–grapheme correspondences). Hence, according to these studies, we expected to find larger deficits in reading comprehension for dyslexic readers from more opaque orthographies, since these orthographies keep less skilled readers mired at the floor of the distribution but propel more skilled readers towards the ceiling (Mcclung & Pearson, [<reflink idref="bib34" id="ref99">34</reflink>]).</p> <p>However, our results showed an almost equivalent deficit in reading comprehension for opaque and transparent orthographies (<emph>d</emph> = 0.614 and <emph>d</emph> = 0.452, respectively) suggesting that the orthographic transparency does not affect reading comprehension, at least in adults. In addition, the surprising results observed in intermediate orthographies do not fit Mcclung and Pearson's ([<reflink idref="bib34" id="ref100">34</reflink>]) proposal. But we note that, due to the imbalance in the number of studies representing each orthographic group with consequences for the reliability of ESs estimates and statistical power, results must be looked at carefully. In our opinion, these different findings about the moderator role of orthographies in reading comprehension deficits in adults with dyslexia need to be further explored with extended samples.</p> <p>Regarding the core skills that support reading and writing, our meta-analysis showed that orthographies have a moderator effect on phonological awareness accuracy. Specifically, adults with dyslexia performed markedly worse than controls in intermediate and opaque orthographies (<emph>d</emph>'s > 1.1), while this phonological awareness deficit was smaller (<emph>d</emph> = 0.663) in dyslexic adults from transparent orthographies. Whether this results from a better compensation or from an a priori lower deficit in individuals with dyslexia in transparent orthographies is unclear.</p> <p>The impact of phonological awareness in reading performance across orthographies was already documented by Ziegler and colleagues in children (Ziegler et al., [<reflink idref="bib67" id="ref101">67</reflink>]), showing that it was stronger in more opaque orthographies. Furthermore, Furnes and Samuelsson ([<reflink idref="bib19" id="ref102">19</reflink>]) investigated the longitudinal predictors of reading and spelling difficulties in children learning to read either Norwegian (more transparent) or English (opaque), and found that phonological awareness diminished as a predictor of reading difficulties in transparent orthographies after the first years in school. These cross-cultural studies with children agree that phonological awareness in more transparent orthographies becomes less associated with reading in older children. To confirm that compensation of phonological deficits occurs in more transparent orthographies, longitudinal cross-cultural studies in adults with dyslexia are needed.</p> <p>Since the phonological awareness variable consisted of tests where the participants had to deal with phonological units of different sizes (syllables, rhymes and phonemes) as well as tasks with different complexities (from the simplest such as judgement or matching to more complex ones such as deleting or manipulating phonological units), we reanalysed this variable by considering the nature of the task and the size of the phonological unit. This analysis allowed us to investigate if adults with a lifelong history of reading disorders had somehow compensated for some level of their phonological processing difficulties. The results indicated a medium-to-large impairment in dyslexia that spans the different phonological awareness measures and persists into adulthood.</p> <p>Moreover, in adults with dyslexia, phonological awareness at the syllable level is not necessarily easier than processing phonemes, and performance seems to depend mainly on the task demands. Tasks requiring the manipulation of phonological units, either syllables or phonemes, pose a greater challenge than tasks requiring simpler processing (such as syllables and phoneme blending). It thus seems that an extended experience with the written code might help adult dyslexics to solve simple tasks involving phonemes.</p> <p>In contrast, the size of the deficit in phonological memory, which is closely tied to phonological processing (de Jong & van der Leij, [<reflink idref="bib26" id="ref103">26</reflink>]), was not moderated by the orthography. Deficits in phonological memory as well as in verbal working memory have been consistently observed in children and adolescents with dyslexia (Swanson, Zheng, & Jerman, [<reflink idref="bib57" id="ref104">57</reflink>]), and our data confirmed that these deficits persist into adulthood. Such verbal memory deficits contribute to reading difficulties by preventing an efficient access to the phonological and executive resources required in reading and writing processes.</p> <p>Another well-known cognitive correlate of reading and writing is RAN and was thus also analysed here. As mentioned previously, this meta-analysis showed a large RAN deficit in adults with dyslexia, extending a recent meta-analysis including mainly dyslexic children (Araújo & Faísca, [<reflink idref="bib1" id="ref105">1</reflink>]). Longitudinal research also demonstrated that rapid naming was, in fact, a factor that differentiated individuals with persisting reading disorders from those with improved reading fluency (Eloranta et al., [<reflink idref="bib16" id="ref106">16</reflink>]). This meta-analysis added that the magnitude of this deficit is equivalent across languages, that is, irrespective of the transparency of the orthography in which participants learned to read. This result adds to the growing evidence that shows a reliable and strong association between RAN and reading ability in a wide array of orthographies and across ages (for a review, see Araújo, Reis, Petersson, & Faísca, [<reflink idref="bib2" id="ref107">2</reflink>]). The observed dissociation between alphabetic and non-alphabetic RAN also corroborates with previous findings: RAN-alphabetic showed a larger ES, supporting the view that RAN-alphabetic makes a major contribution to reading.</p> <p>In addition, our meta-analysis addressed the extent to which dyslexia in adulthood involves weaknesses in other domains beyond reading/writing and phonological processing skills, such as processing speed and other oral language skills. Results showed that adults with dyslexia perform significantly weaker than the controls in tasks emphasising speed, either non-linguistic or linguistic. Large ESs were observed in processing speed tasks (<emph>d</emph> = 0.873) and on serial rapid naming (<emph>d</emph> = 1.295 for RAN-alphabetic and <emph>d</emph> = 0.972 for RAN non-alphabetic). Research has shown that these kind of tasks become increasingly important as literacy development progresses, probably because they are more linked to reading fluency than to single word reading accuracy (Pennington & Lefly, [<reflink idref="bib45" id="ref108">45</reflink>]; Puolakanaho et al., [<reflink idref="bib49" id="ref109">49</reflink>]; Snowling, Gallagher, & Frith, [<reflink idref="bib54" id="ref110">54</reflink>]; Torppa, Lyytinen, Erskine, Eklund, & Lyytinen, [<reflink idref="bib61" id="ref111">61</reflink>]).</p> <p>One possibility is that these deficits contribute to reading fluency problems that are particularly hard to remediate in impaired readers (e.g. Thaler, Ebner, Wimmer, & Landerl, [<reflink idref="bib59" id="ref112">59</reflink>]). The specific relation between speed of processing, rapid naming and reading has been discussed in the literature (e.g. Kail & Hall, [<reflink idref="bib27" id="ref113">27</reflink>]; Papadopoulos, Spanoudis, & Georgiou, [<reflink idref="bib40" id="ref114">40</reflink>]; Powell, Stainthorp, Stuart, Garwood, & Quinlan, [<reflink idref="bib48" id="ref115">48</reflink>]; Vaessen, Gerretsen, & Blomert, [<reflink idref="bib63" id="ref116">63</reflink>]). Some authors (e.g. Kail & Hall, [<reflink idref="bib27" id="ref117">27</reflink>]) have argued that naming speed is related to reading ability because both draw upon a domain-general speed-of-processing factor. Yet, others have shown that, while RAN does share some part of its predictive variance with speed of processing, it still accounts for a significant amount of variance in reading ability beyond the effects of processing speed (e.g. Papadopoulos, Georgiou, & Kendeou, [<reflink idref="bib41" id="ref118">41</reflink>]; Powell et al., [<reflink idref="bib48" id="ref119">48</reflink>]). In a recent study with adult dyslexics (Georgiou et al., [<reflink idref="bib20" id="ref120">20</reflink>]), it was also demonstrated that entering processing speed as a covariate is not sufficient to eliminate the differences in RAN components between dyslexics and their reading controls. Thus, evidence thus far seems to be more compatible with the idea that processing speed is a factor in RAN performance, but it does not account for the relationship between RAN and reading.</p> <p>Regarding the existence of a vocabulary deficit in adults with dyslexia, the literature is not consensual (see, for a review, Cavalli et al., [<reflink idref="bib11" id="ref121">11</reflink>]). Vocabulary was the reading- and writing-associated ability that presented the smallest ES compared with the other skills in this meta-analysis. Yet, the effect was medium (<emph>d</emph> = 0.591), showing that dyslexic adults do, indeed, have a deficit in this skill as already shown in Swanson and Hsieh's ([<reflink idref="bib56" id="ref122">56</reflink>]) meta-analysis. Hence, this result seems to suggest that dyslexics' reading difficulties throughout life represent a disadvantage for vocabulary acquisition. At least in children, vocabulary knowledge has been associated with specific reading skills, particularly with decoding and reading comprehension (Ouellette, [<reflink idref="bib38" id="ref123">38</reflink>]). Interestingly, here we found positive significant associations of vocabulary with word reading speed (<emph>k</emph> = 24, <emph>β</emph> = 0.487, <emph>p</emph> < 0.001) and reading comprehension (<emph>k</emph> = 6, <emph>β</emph> = 0.789, <emph>p</emph> = 0.020). Although our data do not allow conclusions about causality, these results are compatible with the interpretation that reading comprehension difficulties in dyslexia might be partially explained by vocabulary weaknesses. In addition, the same explanation can justify the association between vocabulary and word reading speed, suggesting that poor vocabulary knowledge in adults hinders reading skills.</p> <p>Finally, given the diversity of inclusion and exclusion criteria used in the 178 studies analysed, we considered it pertinent to assess the impact on ESs of two aspects related to the selection procedures of the dyslexic participants, namely requiring or not requiring a previous formal diagnosis and the exclusion of participants with attentional disorders. The results indicate that having a formal diagnosis and controlling for attention deficits do not affect the ESs of the main reading and writing variables.</p> <hd id="AN0146733574-37">Conclusion</hd> <p>In conclusion, our meta-analysis is compatible with the view that the cognitive influences on dyslexia are multifactorial and involve reading and phonological processing deficits as well as weaknesses in other oral language skills (e.g. vocabulary). Critically, these deficits in cognitive processes that underlie poor reading skills persist into adulthood. However, symptoms are more accentuated for reading and writing skills than for reading- and writing-associated processes, such as phonological awareness, rapid automatised naming, phonological memory, verbal working memory and vocabulary. In addition, the primary symptoms associated with developmental dyslexia seem to be amplified when performance is measured with speed. Orthography seems to be an important factor in how symptoms are expressed. Participants who learned to read and write from transparent orthographies showed, in adulthood, less marked symptoms as compared with participants from more opaque orthographies, especially when performance was assessed with accuracy measures. Yet, deficits are larger and more homogeneous across the three orthographies when speed measures are used. Phonological awareness (accuracy) seems to be a minor problem in adulthood mainly for transparent orthographies. We must mention that most of the studies on adults with dyslexia use samples of university students, including this review, which means that their findings come from potentially compensated dyslexics.</p> <hd id="AN0146733574-38">Funding</hd> <p>This work was supported by the Portuguese Foundation for Science and Technology—FCT (UID/BIM/04773/2019 CBMR, PTDC/MHC-PCN/1175/2014 and IF/00533/2015).</p> <hd id="AN0146733574-39">Electronic supplementary material</hd> <p>Graph: (PDF 454kb)</p> <hd id="AN0146733574-40">Publisher's note</hd> <p>Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.</p> <ref id="AN0146733574-41"> <title> References </title> <blist> <bibl id="bib1" idref="ref1" type="bt">1</bibl> <bibtext> Araújo S, Faísca L. A meta-analytic review of naming-speed deficits in developmental dyslexia. Scientific Studies of Reading. 2019; 23; 5: 349-368. 10.1080/10888438.2019.1572758</bibtext> </blist> <blist> <bibl id="bib2" idref="ref2" type="bt">2</bibl> <bibtext> Araújo S, Reis A, Petersson KM, Faísca L. Rapid automatized naming and reading performance: A meta-analysis. Journal of Educational Psychology. 2015; 107; 3: 868-883. 10.1037/edu0000006</bibtext> </blist> <blist> <bibl id="bib3" idref="ref3" type="bt">3</bibl> <bibtext> Becker J, Czamara D, Scerri TS, Ramus F, Csépe V, Talcott JB, Honbolygó F. Genetic analysis of dyslexia candidate genes in the European cross-linguistic NeuroDys cohort. European Journal of Human Genetics. 2014; 22; 5: 675-680. 10.1038/ejhg.2013.199</bibtext> </blist> <blist> <bibl id="bib4" idref="ref41" type="bt">4</bibl> <bibtext> Borenstein, M, Hedges, L, Higgins, J, & Rothstein, H. (2005). Comprehensive meta-analysis (version 2). Englewood, NJ: Biostat, 104.</bibtext> </blist> <blist> <bibl id="bib5" idref="ref47" type="bt">5</bibl> <bibtext> Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Introduction to meta-analysis. 2009: Chichester, England; Wiley</bibtext> </blist> <blist> <bibl id="bib6" idref="ref94" type="bt">6</bibl> <bibtext> Bosman A, Van Orden GPerfetti C, Rieben L, Fayol M. Why spelling is more difficult than reading. Learning to spell: Research, theory, and practice across languages. 1997: Hillsdale, N.J; Erlbaum: 173-194</bibtext> </blist> <blist> <bibl id="bib7" idref="ref88" type="bt">7</bibl> <bibtext> Breznitz Z. Effects of accelerated reading rate on memory for text among dyslexic readers. Journal of Educational Psychology. 1997; 89; 2: 289-297. 10.1037/0022-0663.89.2.289</bibtext> </blist> <blist> <bibl id="bib8" idref="ref14" type="bt">8</bibl> <bibtext> Callens M, Tops W, Stevens M, Brysbaert M. An exploratory factor analysis of the cognitive functioning of first-year bachelor students with dyslexia. Annals of Dyslexia. 2014; 64; 1: 91-119. 10.1007/s11881-013-0088-6</bibtext> </blist> <blist> <bibl id="bib9" idref="ref62" type="bt">9</bibl> <bibtext> Caravolas M, Lervåg A, Defior S, Seidlová Málková G, Hulme C. Different patterns, but equivalent predictors, of growth in reading in consistent and inconsistent orthographies. Psychological Science. 2013; 24; 8: 1398-1407. 10.1177/0956797612473122</bibtext> </blist> <blist> <bibtext> Cavalli E, Duncan LG, Elbro C, El Ahmadi A, Colé P. Phonemic-morphemic dissociation in university students with dyslexia: An index of reading compensation?. Annals of Dyslexia. 2017; 67; 1: 63-84. 10.1007/s11881-016-0138-y</bibtext> </blist> <blist> <bibtext> Cavalli, E, Casalis, S, Ahmadi, A, Zira, M, Poracchia-George, F, & Colé, P. (2016). Vocabulary skills are well developed in university students with dyslexia: Evidence from multiple case studies. Research in Developmental Disabilities, 51-52, 89–102. 10.1016/j.ridd.2016.01.006.</bibtext> </blist> <blist> <bibtext> Cohen, J. (1988). Statistical power analysis for the behavioral sciences. 2nd ed. Lawrence Erlbaum Associates.</bibtext> </blist> <blist> <bibtext> Cooper HM. Research synthesis and meta-analysis: A step-by-step approach. 20104: Thousand Oaks, CA; Sage</bibtext> </blist> <blist> <bibtext> Denckla MB, Rudel RG. Rapid "Automatized" Naming (R.A.N.): Dyslexia differentiated from other learning disabilities. Neuropsychologia. 1976; 14: 471-479. 10.1016/0028-3932(76)90075-0</bibtext> </blist> <blist> <bibtext> Duval, S, & Tweedie, R. (2000). Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta‐analysis. Biometrics, 56(2), 455–463.</bibtext> </blist> <blist> <bibtext> Eloranta AK, Närhi VM, Eklund KM, Ahonen TPS, Aro TI. Resolving reading disability—Childhood predictors and adult-age outcomes. Dyslexia. 2019; 25; 1: 20-37. 10.1002/dys.1605</bibtext> </blist> <blist> <bibtext> Everatt J. The abilities and disabilities associated with adult developmental dyslexia. Journal of Research in Reading. 1997; 20; 1: 13-21. 10.1111/1467-9817.00016</bibtext> </blist> <blist> <bibtext> Ferrer E, Shaywitz BA, Holahan JM, Marchione K, Shaywitz SE. Uncoupling of reading and IQ over time: Empirical evidence for a definition of dyslexia. Psychological Science. 2010; 21; 1: 93-101. 10.1177/0956797609354084</bibtext> </blist> <blist> <bibtext> Furnes, B, & Samuelsson, S. (2011). Predicting reading and spelling difficulties in transparent and opaque orthographies: A comparison between Scandinavian and U.S./Australian Children. Dyslexia, 16(2), 119–142. 10.1002/dys.401.</bibtext> </blist> <blist> <bibtext> Georgiou GK, Ghazyani R, Parrila R. Are RAN deficits in university students with dyslexia due to defective lexical access, impaired anchoring, or slow articulation?. Annals of Dyslexia. 2018; 68: 85-103</bibtext> </blist> <blist> <bibtext> Gerber PJ. The impact of learning disabilities on adulthood: A review of the evidenced-based literature for research and practice in adult education. Journal of Learning Disabilities. 2012; 45; 1: 31-46. 10.1177/0022219411426858</bibtext> </blist> <blist> <bibtext> Goulandris N. Dyslexia in different languages: Cross-linguistic comparisons. 2003: London; Whurr</bibtext> </blist> <blist> <bibtext> Hanley R, Masterson J, Spencer L, Evans D. How long do the advantages of learning to read a transparent orthography last? An investigation of the reading skills and reading impairment of Welsh children at 10 years of age. The Quarterly Journal of Experimental Psychology Section A: Human Experimental Psychology. 2004; 57A; 6: 1393-1410. 10.1080/02724980343000819</bibtext> </blist> <blist> <bibtext> Hatcher J, Snowling MJ, Griffiths YM. Cognitive assessment of dyslexic students in higher education. British Journal of Educational Psychology. 2002; 72; 1: 119-133. 10.1348/000709902158801</bibtext> </blist> <blist> <bibtext> Hedges LV. Distribution theory for Glass's estimator of effect size and related estimators. Journal of Educational Statistics. 1981; 6; 2: 107-128. 10.3102/10769986006002107</bibtext> </blist> <blist> <bibtext> de Jong PF, van der Leij A. Specific contributions of phonological abilities to early reading acquisition: Results from a Dutch latent variable longitudinal study. Journal of Educational Psychology. 1999; 91; 3: 450-476. 10.1037/0022-0663.91.3.450</bibtext> </blist> <blist> <bibtext> Kail R, Hall LK. Processing speed, naming speed, and reading. Developmental Psychology. 1994; 30; 6: 949-954. 10.1037/0012-1649.30.6.949</bibtext> </blist> <blist> <bibtext> Kemp N, Parrila RK, Kirby JR. Phonological and orthographic spelling in high-functioning adult dyslexics. Dyslexia. 2009; 15: 105-128. 10.1002/dys.364</bibtext> </blist> <blist> <bibtext> Landerl K, Freudenthaler HH, Heene M, De Jong PF, Desrochers A, Manolitsis G, Georgiou GK. Phonological awareness and rapid automatized naming as longitudinal predictors of reading in five alphabetic orthographies with varying degrees of consistency. Scientific Studies of Reading. 2019; 23; 3: 220-234. 10.1080/10888438.2018.1510936</bibtext> </blist> <blist> <bibtext> Landerl K, Ramus F, Moll K, Lyytinen H, Leppänen PHT, Lohvansuu K, Kunze S. Predictors of developmental dyslexia in European orthographies with varying complexity. Journal of Child Psychology and Psychiatry. 2013; 54; 6: 686-694. 10.1111/jcpp.12029</bibtext> </blist> <blist> <bibtext> Lindgrén SA, Laine M. Cognitive-linguistic performances of multilingual university students suspected of dyslexia. Dyslexia. 2011; 17; 2: 184-200. 10.1002/dys.422</bibtext> </blist> <blist> <bibtext> Lefly DL, Pennington BF. Spelling errors and reading fluency in compensated adult dyslexics. Annals of Dyslexia. 1991; 41; 1: 143-162. 10.1007/BF02648083</bibtext> </blist> <blist> <bibtext> Lyytinen H, Aro M, Holopainnen LSmythe I, Everatt J, Slater R. Dyslexia in highly ortographically regular finish. International book of dyslexia: A cross language comparison and practice guide. 2004: Chichester, England; Wiley: 81-91</bibtext> </blist> <blist> <bibtext> Mcclung NA, Pearson PD. Reading comprehension across languages: Seven European orthographies and two international literacy assessments. Written Language & Literacy. 2019; 22; 1: 33-66. 10.1075/wll.00019.mcc</bibtext> </blist> <blist> <bibtext> Miller-Shaul S. The characteristics of young and adult dyslexics readers on reading and reading related cognitive tasks as compared to normal readers. Dyslexia. 2005; 11: 132-151. 10.1002/dys.290</bibtext> </blist> <blist> <bibtext> Moll K, Ramus F, Bartling J, Bruder J, Kunze S, Neuhoff N, Tóth D. Cognitive mechanisms underlying reading and spelling development in five European orthographies. Learning and Instruction. 2014; 29: 65-77. 10.1016/j.learninstruc.2013.09.003</bibtext> </blist> <blist> <bibtext> Nergård-Nilssen T, Hulme C. Developmental dyslexia in adults: Behavioural manifestations and cognitive correlates. Dyslexia. 2014; 20; 3: 191-207. 10.1002/dys.1477</bibtext> </blist> <blist> <bibtext> Ouellette GP. What's meaning got to do with it: The role of vocabulary in word reading and reading comprehension. Journal of Educational Psychology. 2006; 98; 3: 554-566. 10.1037/0022-0663.98.3.554</bibtext> </blist> <blist> <bibtext> Pammer K. Brain mechanisms and reading remediation: More questions than answers. Scientifica. 2014; 2014: 1-9. 10.1155/2014/802741</bibtext> </blist> <blist> <bibtext> Papadopoulos TC, Spanoudis GC, Georgiou GK. How is RAN related to reading fluency? A comprehensive examination of the prominent theoretical accounts. Frontiers in Psychology. 2016; 7; Aug: 1-15. 10.3389/fpsyg.2016.01217</bibtext> </blist> <blist> <bibtext> Papadopoulos TC, Georgiou GK, Kendeou P. Investigating the double-deficit hypothesis in Greek: Findings from a longitudinal study. Journal of Learning Disabilities. 2009; 42; 6: 528-547. 10.1177/0022219409338745</bibtext> </blist> <blist> <bibtext> Parrila R, Dudley D, Song S, Georgiou G. A meta-analysis of reading-level match dyslexia studies in consistent alphabetic orthographies. Annals of Dyslexia. 2020; 70: 1-26. 10.1007/s11881-019-00187-5A</bibtext> </blist> <blist> <bibtext> Paulesu E, Démonet J-FF, Fazio F, McCrory E, Chanoine V, Brunswick N, Frith U. Dyslexia: Cultural diversity and biological unity. Science. 2001; 291; 5511: 2165-2167. 10.1126/science.1057179</bibtext> </blist> <blist> <bibtext> Pennington BF, van Orden GC, Smith SD, Green PA, Haith MM. Phonological processing skills and deficits in adult dyslexics. Child Development. 1990; 61; 1: 1753-1778. 10.2307/1130836</bibtext> </blist> <blist> <bibtext> Pennington BF, Lefly DL. Early reading development in children at family risk for dyslexia. Child Development. 2001; 72; 3: 816-833. 10.1111/1467-8624.00317</bibtext> </blist> <blist> <bibtext> Perfetti C, Stafura J. Word knowledge in a theory of reading comprehension. Scientific Studies of Reading. 2014; 18; 1: 22-37. 10.1080/10888438.2013.827687</bibtext> </blist> <blist> <bibtext> Pino M, Mortari L. The inclusion of students with dyslexia in higher education: A systematic review using narrative synthesis. Dyslexia. 2014; 20; 4: 346-369. 10.1002/dys.1484</bibtext> </blist> <blist> <bibtext> Powell D, Stainthorp R, Stuart M, Garwood H, Quinlan P. An experimental comparison between rival theories of rapid automatized naming performance and its relationship to reading. Journal of Experimental Child Psychology. 2007; 98: 46-68. 10.1016/j.jecp.2007.04.003</bibtext> </blist> <blist> <bibtext> Puolakanaho A, Ahonrn T, Aro M, Eklund K, Leppänen PHT, Poikkeus A-M, Lyytinen H. Developmental links of very early phonological and language skills to second grade reading outcomes: Strong to accuracy but only minor to fluency. Journal of Learning Disabilities. 2008; 41; 4: 353-370. 10.1177/0022219407311747</bibtext> </blist> <blist> <bibtext> Re AM, Tressoldi PE, Cornoldi C, Lucangeli D. Which tasks best discriminate between dyslexic university students and controls in a transparent language?. Dyslexia. 2011; 17: 227-241. 10.1002/dys.431</bibtext> </blist> <blist> <bibtext> Rosenthal R. Meta-analytic procedures for social science research. 1991: London, England; Sage</bibtext> </blist> <blist> <bibtext> Seymour PHK, Aro M, Erskine JM. Foundation literacy acquisition in European orthographies. British Journal of Psychology. 2003; 94: 143-174. 10.1348/000712603321661859</bibtext> </blist> <blist> <bibtext> Share DL. On the Anglocentricities of current reading research and practice: The perils of overreliance on an "outlier" orthography. Psychological Bulletin. 2008; 134; 4: 584-615. 10.1037/0033-2909.134.4.584</bibtext> </blist> <blist> <bibtext> Snowling MJ, Gallagher A, Frith U. Family risk of dyslexia is continuous: Individual differences in the precursors of reading skill. Child Development. 2003; 74; 2: 358-373</bibtext> </blist> <blist> <bibtext> Suárez-Coalla P, Cuetos F. Reading difficulties in Spanish adults with dyslexia. Annals of Dyslexia. 2015; 65; 1: 33-51. 10.1007/s11881-015-0101-3</bibtext> </blist> <blist> <bibtext> Swanson HL, Hsieh C-J. Reading disabilities in adults: A selective meta-analysis of the literature. Review of Educational Research. 2009; 79; 4: 1362-1390. 10.3102/0034654309350931</bibtext> </blist> <blist> <bibtext> Swanson HL, Zheng X, Jerman O. Working memory, short-term memory, and reading disabilities. A selective meta-analysis of the literature. Journal of Learning Disabilities. 2009; 42; 3: 260-297</bibtext> </blist> <blist> <bibtext> Tamboer P, Vorst HCM, Oort FJ. Five describing factors of dyslexia. Journal of Learning Disabilities. 2016; 49: 451-465. 10.1177/0022219414558123</bibtext> </blist> <blist> <bibtext> Thaler V, Ebner EM, Wimmer H, Landerl K. Training reading fluency in dysfluent readers with high reading accuracy: Word specific effects but low transfer to untrained words. Annals of Dyslexia. 2004; 54; 1: 89-113. 10.1007/s11881-004-0005-0</bibtext> </blist> <blist> <bibtext> Tops W, Callens M, Lammertyn J, van Hees V, Brysbaert M. Identifying students with dyslexia in higher education. Annals of Dyslexia. 2012; 62; 3: 186-203. 10.1007/s11881-012-0072-6</bibtext> </blist> <blist> <bibtext> Torppa M, Lyytinen P, Erskine J, Eklund K, Lyytinen H. Language development, literacy skills, and predictive connections to reading in Finnish children with and without familial risk for dyslexia. Jornal of Learning Disabilities. 2010; 43; 4: 308-321. 10.1177/0022219410369096</bibtext> </blist> <blist> <bibtext> Vaessen A, Bertrand D, Tóth D, Csépe V, Faísca L, Reis A, Blomert L. Cognitive development of fluent word reading does not qualitatively differ between transparent and opaque orthographies. Journal of Educational Psychology. 2010; 102; 4: 827-842. 10.1037/a0019465</bibtext> </blist> <blist> <bibtext> Vaessen A, Gerretsen P, Blomert L. Naming problems do not reflect a second independent core deficit in dyslexia: Double deficits explored. Journal of Experimental Child Psychology. 2009; 103; 2: 202-221. 10.1016/j.jecp.2008.12.004</bibtext> </blist> <blist> <bibtext> Verhoeven, L, & Perfetti, C. (2017). In L. Verhoeven & C. Perfetti (Eds.), Learning to read across languages and writing systems. Cambridge: Cambridge University Press. 10.1017/9781316155752.</bibtext> </blist> <blist> <bibtext> Watson J, Boman P. Mainstreamed students with learning difficulties: Failing and underachieving in the secondary school. Australian Journal of Learning Difficulties. 2005; 10; 2: 43-49. 10.1080/19404150509546788</bibtext> </blist> <blist> <bibtext> Yap R, van der Leij A. Word processing in dyslexics: An automatic decoding deficit?. Reading and Writing. 1993; 5: 261-279</bibtext> </blist> <blist> <bibtext> Ziegler JC, Bertrand D, Toth D, Csepe V, Reis A, Faísca L, Blomert L. Orthographic depth and its impact on universal predictors of reading: A cross-language investigation. Psychological Science. 2010; 21; 4: 551-559. 10.1177/0956797610363406</bibtext> </blist> <blist> <bibtext> Ziegler JC, Perry C, Ma-Wyatt A, Ladner D, Schulte-Körne G. Developmental dyslexia in different languages: Language-specific or universal?. Journal of Experimental Child Psychology. 2003; 86; 3: 169-193. 10.1016/S0022-0965(03)00139-5</bibtext> </blist> <blist> <bibtext> Zoccolotti P, De Luca M, Pace E, Udica A, Orlandi M, Spinelli D. Markers of developmental surface dyslexia in a language (Italian) with high grapheme-phoneme correspondence. Applied PsychoLinguistics. 1999; 20; 2: 191-216</bibtext> </blist> </ref> <aug> <p>By Alexandra Reis; Susana Araújo; Inês Salomé Morais and Luís Faísca</p> <p>Reported by Author; Author; Author; Author</p> </aug> <nolink nlid="nl1" bibid="bib24" firstref="ref4"></nolink> <nolink nlid="nl2" bibid="bib39" firstref="ref5"></nolink> <nolink nlid="nl3" bibid="bib21" firstref="ref6"></nolink> <nolink nlid="nl4" bibid="bib65" firstref="ref7"></nolink> <nolink nlid="nl5" bibid="bib35" firstref="ref8"></nolink> <nolink nlid="nl6" bibid="bib32" firstref="ref9"></nolink> <nolink nlid="nl7" bibid="bib10" firstref="ref10"></nolink> <nolink nlid="nl8" bibid="bib16" firstref="ref11"></nolink> <nolink nlid="nl9" bibid="bib47" firstref="ref13"></nolink> <nolink nlid="nl10" bibid="bib56" firstref="ref15"></nolink> <nolink nlid="nl11" bibid="bib52" firstref="ref16"></nolink> <nolink nlid="nl12" bibid="bib29" firstref="ref17"></nolink> <nolink nlid="nl13" bibid="bib62" firstref="ref18"></nolink> <nolink nlid="nl14" bibid="bib67" firstref="ref19"></nolink> <nolink nlid="nl15" bibid="bib30" firstref="ref20"></nolink> <nolink nlid="nl16" bibid="bib68" firstref="ref21"></nolink> <nolink nlid="nl17" bibid="bib50" firstref="ref22"></nolink> <nolink nlid="nl18" bibid="bib15" firstref="ref23"></nolink> <nolink nlid="nl19" bibid="bib44" firstref="ref24"></nolink> <nolink nlid="nl20" bibid="bib37" firstref="ref25"></nolink> <nolink nlid="nl21" bibid="bib60" firstref="ref26"></nolink> <nolink nlid="nl22" bibid="bib53" firstref="ref28"></nolink> <nolink nlid="nl23" bibid="bib42" firstref="ref35"></nolink> <nolink nlid="nl24" bibid="bib14" firstref="ref36"></nolink> <nolink nlid="nl25" bibid="bib64" firstref="ref40"></nolink> <nolink nlid="nl26" bibid="bib25" firstref="ref42"></nolink> <nolink nlid="nl27" bibid="bib51" firstref="ref43"></nolink> <nolink nlid="nl28" bibid="bib12" firstref="ref44"></nolink> <nolink nlid="nl29" bibid="bib13" firstref="ref45"></nolink> <nolink nlid="nl30" bibid="bib160" firstref="ref49"></nolink> <nolink nlid="nl31" bibid="bib135" firstref="ref50"></nolink> <nolink nlid="nl32" bibid="bib38" firstref="ref51"></nolink> <nolink nlid="nl33" bibid="bib36" firstref="ref52"></nolink> <nolink nlid="nl34" bibid="bib98" firstref="ref53"></nolink> <nolink nlid="nl35" bibid="bib100" firstref="ref54"></nolink> <nolink nlid="nl36" bibid="bib76" firstref="ref59"></nolink> <nolink nlid="nl37" bibid="bib27" firstref="ref60"></nolink> <nolink nlid="nl38" bibid="bib74" firstref="ref61"></nolink> <nolink nlid="nl39" bibid="bib58" firstref="ref66"></nolink> <nolink nlid="nl40" bibid="bib18" firstref="ref71"></nolink> <nolink nlid="nl41" bibid="bib22" firstref="ref80"></nolink> <nolink nlid="nl42" bibid="bib43" firstref="ref81"></nolink> <nolink nlid="nl43" bibid="bib33" firstref="ref85"></nolink> <nolink nlid="nl44" bibid="bib69" firstref="ref86"></nolink> <nolink nlid="nl45" bibid="bib66" firstref="ref87"></nolink> <nolink nlid="nl46" bibid="bib17" firstref="ref90"></nolink> <nolink nlid="nl47" bibid="bib28" firstref="ref91"></nolink> <nolink nlid="nl48" bibid="bib31" firstref="ref93"></nolink> <nolink nlid="nl49" bibid="bib23" firstref="ref96"></nolink> <nolink nlid="nl50" bibid="bib34" firstref="ref97"></nolink> <nolink nlid="nl51" bibid="bib46" firstref="ref98"></nolink> <nolink nlid="nl52" bibid="bib19" firstref="ref102"></nolink> <nolink nlid="nl53" bibid="bib26" firstref="ref103"></nolink> <nolink nlid="nl54" bibid="bib57" firstref="ref104"></nolink> <nolink nlid="nl55" bibid="bib45" firstref="ref108"></nolink> <nolink nlid="nl56" bibid="bib49" firstref="ref109"></nolink> <nolink nlid="nl57" bibid="bib54" firstref="ref110"></nolink> <nolink nlid="nl58" bibid="bib61" firstref="ref111"></nolink> <nolink nlid="nl59" bibid="bib59" firstref="ref112"></nolink> <nolink nlid="nl60" bibid="bib40" firstref="ref114"></nolink> <nolink nlid="nl61" bibid="bib48" firstref="ref115"></nolink> <nolink nlid="nl62" bibid="bib63" firstref="ref116"></nolink> <nolink nlid="nl63" bibid="bib41" firstref="ref118"></nolink> <nolink nlid="nl64" bibid="bib20" firstref="ref120"></nolink> <nolink nlid="nl65" bibid="bib11" firstref="ref121"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1275867
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Reading and Reading-Related Skills in Adults with Dyslexia from Different Orthographic Systems: A Review and Meta-Analysis
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Reis%2C+Alexandra%22">Reis, Alexandra</searchLink> (ORCID <externalLink term="http://orcid.org/0000-0001-5598-0999">0000-0001-5598-0999</externalLink>)<br /><searchLink fieldCode="AR" term="%22Araújo%2C+Susana%22">Araújo, Susana</searchLink><br /><searchLink fieldCode="AR" term="%22Morais%2C+Inês+Salomé%22">Morais, Inês Salomé</searchLink><br /><searchLink fieldCode="AR" term="%22Faísca%2C+Luís%22">Faísca, Luís</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Annals+of+Dyslexia%22"><i>Annals of Dyslexia</i></searchLink>. Oct 2020 70(3):339-368.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Springer. Available from: Springer Nature. One New York Plaza, Suite 4600, New York, NY 10004. Tel: 800-777-4643; Tel: 212-460-1500; Fax: 212-460-1700; e-mail: customerservice@springernature.com; Web site: https://link.springer.com/
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 30
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2020
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Reading+Skills%22">Reading Skills</searchLink><br /><searchLink fieldCode="DE" term="%22Adults%22">Adults</searchLink><br /><searchLink fieldCode="DE" term="%22Dyslexia%22">Dyslexia</searchLink><br /><searchLink fieldCode="DE" term="%22Orthographic+Symbols%22">Orthographic Symbols</searchLink><br /><searchLink fieldCode="DE" term="%22Reading+Difficulties%22">Reading Difficulties</searchLink><br /><searchLink fieldCode="DE" term="%22Reading+Comprehension%22">Reading Comprehension</searchLink><br /><searchLink fieldCode="DE" term="%22Reading+Rate%22">Reading Rate</searchLink><br /><searchLink fieldCode="DE" term="%22Phonological+Awareness%22">Phonological Awareness</searchLink><br /><searchLink fieldCode="DE" term="%22Writing+Difficulties%22">Writing Difficulties</searchLink><br /><searchLink fieldCode="DE" term="%22Symptoms+%28Individual+Disorders%29%22">Symptoms (Individual Disorders)</searchLink><br /><searchLink fieldCode="DE" term="%22Severity+%28of+Disability%29%22">Severity (of Disability)</searchLink><br /><searchLink fieldCode="DE" term="%22Cognitive+Processes%22">Cognitive Processes</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1007/s11881-020-00205-x
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0736-9387
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: An individual diagnosed with dyslexia in childhood typically remains dyslexic throughout his/her life. However, the cognitive profile of adults with dyslexia has been less explored than that of children. This meta-analytic study is intended to clarify three questions: (1) To what extent, and in what manner, do adults with reading difficulties (dyslexia) differ from typical adult readers in measures of reading and writing competence and related cognitive skills?; (2) To what extent do speed measures pose a greater challenge than accuracy measures in an adult population that has already had years of print exposure?; and (3) To what extent does orthographic transparency modulate the reading profile of adults with dyslexia? A total of 178 studies comparing adults with dyslexia and matched controls were reviewed. The results showed that adults with dyslexia exhibited poor performance on almost all reading and writing tasks expressed by very large effect sizes (range 1.735 [less than or equal to] "d" [less than or equal to] 2.034), except for reading comprehension ("d" = 0.729). Deficits in reading- and writing-related variables are also present but with a lower expression (range 0.591 [less than or equal to] "d" [less than or equal to] 1.295). These difficulties are exacerbated for speed measures, especially for word and pseudoword reading, phonological awareness and orthographic knowledge. Orthographic transparency proved to be a significant moderator of dyslexic deficits in word and pseudoword reading, reading comprehension, spelling and phonological awareness, with the expression of the deficits being weaker on transparent--as opposed to intermediate and opaque--orthographies. Overall, the meta-analysis shows that reading and writing difficulties persist in adulthood and are more pronounced in speed measures. Moreover, symptoms are more severe for reading and writing than they are for measures tapping into the cognitive processes underlying reading skills. Orthographic transparency has a significant effect on the manifestation of dyslexia, with dyslexia symptoms being less marked on transparent orthographies. In addition, phonological awareness seems to be a minor problem in adulthood, especially for transparent orthographies.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2020
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1275867
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1275867
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1007/s11881-020-00205-x
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 30
        StartPage: 339
    Subjects:
      – SubjectFull: Reading Skills
        Type: general
      – SubjectFull: Adults
        Type: general
      – SubjectFull: Dyslexia
        Type: general
      – SubjectFull: Orthographic Symbols
        Type: general
      – SubjectFull: Reading Difficulties
        Type: general
      – SubjectFull: Reading Comprehension
        Type: general
      – SubjectFull: Reading Rate
        Type: general
      – SubjectFull: Phonological Awareness
        Type: general
      – SubjectFull: Writing Difficulties
        Type: general
      – SubjectFull: Symptoms (Individual Disorders)
        Type: general
      – SubjectFull: Severity (of Disability)
        Type: general
      – SubjectFull: Cognitive Processes
        Type: general
    Titles:
      – TitleFull: Reading and Reading-Related Skills in Adults with Dyslexia from Different Orthographic Systems: A Review and Meta-Analysis
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Reis, Alexandra
      – PersonEntity:
          Name:
            NameFull: Araújo, Susana
      – PersonEntity:
          Name:
            NameFull: Morais, Inês Salomé
      – PersonEntity:
          Name:
            NameFull: Faísca, Luís
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 10
              Type: published
              Y: 2020
          Identifiers:
            – Type: issn-print
              Value: 0736-9387
          Numbering:
            – Type: volume
              Value: 70
            – Type: issue
              Value: 3
          Titles:
            – TitleFull: Annals of Dyslexia
              Type: main
ResultId 1