Professional Documents
Culture Documents
net/publication/353938612
CITATIONS READS
12 785
9 authors, including:
Peng Peng
University of Texas at Austin
97 PUBLICATIONS 2,824 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
SRCD 2021 - Paper Symposium: Leveraging Innovative Methods in Meta-Analysis to Advance Developmental Theory View project
All content following this page was uploaded by Xiujie Yang on 01 November 2021.
CITATION
Yang, X., Yan, M., Ruan, Y., Ku, S. Y. Y., Lo, J. C. M., Peng, P., & McBride, C. (2021, October 28). Relations Among
Phonological Processing Skills and Mathematics in Children: A Meta-Analysis. Journal of Educational Psychology. Advance
online publication. http://dx.doi.org/10.1037/edu0000710
Journal of Educational Psychology
© 2021 American Psychological Association
ISSN: 0022-0663 https://doi.org/10.1037/edu0000710
Xiujie Yang1, Mengge Yan2, Yijun Ruan2, Serena Yuk Yee Ku2, Jason Chor Ming Lo3, Peng Peng4,
and Catherine McBride2
1
Faculty of Psychology, Beijing Normal University
2
Department of Psychology, The Chinese University of Hong Kong
3
Department of Pediatrics, Haskins Laboratories, New Haven, Connecticut, United States
4
Department of Special Education, University of Texas at Austin
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
The present study presents a meta-analysis of the relations between phonological processing abilities
and different mathematics subskills. Using a random-effects model with 94 studies (135 unique samples,
826 effect sizes), the present meta-analysis revealed a significant general association between phonolog-
ical processing and mathematics (average r = .33, p , .001, 95% CI [.30, .36]). Phonological awareness
(PA) and rapid automatized naming (RAN) showed stronger correlations with mathematics than phono-
logical memory (PM) did. The correlations among phonological processing abilities and mathematics
skills were generally stronger among younger children than among older children. PA and PM mani-
fested larger effect sizes when correlated with mathematics accuracy than with mathematics fluency,
whereas RAN yielded larger effect sizes when associated with mathematics fluency than with mathe-
matics accuracy. Metastructural equation modeling results revealed that, after statistically controlling for
domain-general abilities (i.e., vocabulary knowledge, executive functioning, and nonverbal intelligence),
phonological processing still made a unique contribution to different mathematics subskills (bs = .20
.54). These results suggest that children may use phonological processing abilities as one mechanism
through which to represent, manipulate, and retrieve mathematics knowledge.
1
2 YANG ET AL.
Phonological processing refers to the use of the sound structure will allow us to validate the mechanism of phonological processing
of language to process written and oral information (Liberman et abilities and their relations with the subskills of mathematics. Practi-
al., 1974; Mann & Liberman, 1984; McBride-Chang & Ho, 2000; cally, phonological processing skills emerge before children enter
Wagner, 1986; Wagner & Torgesen, 1987). It includes three com- school, and early assessments of these skills might facilitate identifi-
ponents, namely, phonological awareness (hereafter, PA), rapid cation of children at risk of academic failure. For example, research-
automatized naming (hereafter, RAN), and phonological memory ers may rely on nondigit RAN and nondigit PM as predictors of
(hereafter, PM; Turkeltaub et al., 2003; Wagner et al., 1987, 1994; mathematics performance even before children go to school and
Wagner & Torgesen, 1987). become familiar with digits. Moreover, understanding which phono-
In the past decades, researchers have suggested various roles of logical processing component(s) is/are essential for different mathe-
phonological processing abilities in accounting for mathematics matics abilities could help develop interventions incorporating
skills (e.g., Hecht et al., 2001; Simmons & Singleton, 2008). For phonological processing for children with or without mathematics
example, in the pathway model by LeFevre et al. (2010), phono- disabilities.
logical processing skills are important components of the linguistic
framework that contributes to early mathematics development, Phonological Awareness
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Smedt et al., 2010; Yang & McBride, 2020). This is likely due to The first of these bodies of research centers on PA, which is
the fact that the development of mathematics knowledge relies awareness of phonology, or the sound structure of one’s language
highly on the encoding and maintenance of accurate phonological (McBride, 2016). How PA is measured varies across ages and lan-
representations of numbers, terms, and operators (Krajewski & guages. This is because different aspects of PA are influenced by
Schneider, 2009; Peng et al., 2020). Mathematics knowledge also age as well as educational experiences. For instance, tone discrimi-
highly relies on the automaticity of arithmetic-fact retrieval (Hecht nation and syllable deletion can be measured even in kindergarten,
et al., 2001) such that arithmetic problem solving demands the re- while phoneme deletion cannot be measured until primary school
trieval of phonological forms corresponding to the number of in Chinese children (Shu et al., 2008). Similarly, syllable-level
words and operation facts stored in long-term memory (De Smedt awareness seems to emerge naturally with development in English,
et al., 2010; Geary, 1995; Geary et al., 1993). Not surprisingly, but phoneme-level awareness is at least partly attributable to liter-
weak phonological representations can impede children’s abilities acy education (Cho & McBride-Chang, 2005; Morrison et al.,
to solve mathematics problems, especially those involving the 1995). The effect of PA on arithmetic remains uncertain given
manipulations of number verbal codes (Simmons & Singleton, some controversial results (e.g., Amland et al., 2021; Cirino et al.,
2008). 2018; De Smedt et al., 2010). For example, Amland et al. (2021)
Koponen et al. (2017) reviewed the relationship between one of revealed that phonological awareness failed to explain the devel-
the phonological components, RAN, and mathematics. They found opment of arithmetic from kindergarten to first grade. De Smedt
a significant association between RAN and mathematics (r = .37; and Boets (2010) investigated whether fact retrieval difficulties in
95% CI [.33, .42], and RAN had a stronger correlation with mathe- arithmetic could be explained by individual differences in the three
matics fluency tasks than with mathematics accuracy tasks. A classic areas of phonological processing, that is, PA, RAN, and
meta-analysis by Friso-Van den Bos et al. (2013) demonstrated a PM, and found that phonological processing, particularly PA, was
medium-sized significant correlation between PM and a number of related to arithmetic fact retrieval.
mathematics domains, with no distinction between subcomponents
within PM (e.g., digit PM, and nondigit PM). However, these stud- Rapid Automatized Naming
ies only involved one or two of the phonological processing skills,
without systematically comparing each phonological processing The second body of phonological processing research focuses
regarding their relations to various mathematics skills. Peng et al. on RAN, which is defined as rate of access to phonological infor-
(2020) investigated the relation between language and mathemat- mation stored in long-term memory (Wagner & Torgesen, 1987).
ics, in which language was indexed with but not limited to phono- RAN is typically measured by timing how quickly individuals can
logical processing and comprehensive language. And it is yet to be name symbol stimuli, both digits and/or nondigits, including colors,
determined whether phonological processing correlate to mathe- objects, or letters. Recent studies have argued that RAN measured
matics after strictly controlling for more domain-general abilities using these stimuli operates at the grapho-phonemic, grapho-syl-
such as nonverbal intelligence, executive functioning, and vocabu- labic, and lexical reading levels (Moll et al., 2009; Onochie-Quin-
lary knowledge. This was the goal of the present study. tanilla et al., 2019). One study demonstrated that RAN, rather
Addressing these gaps is of theoretical and practical importance. than PA, was uniquely associated when predicting arithmetic
The aim of the study was to examine the comparative correlation skills of primary school children who tended to employ fact re-
strengths between various phonological processing skills and differ- trieval strategies (Yang et al., 2019). Other studies have proposed
ent mathematics skills; we intended to investigate material/domain- that RAN is related to the efficient retrieval of phonological name
specific effects by possibly differentiating subcomponents both codes, counting sequencing, and arithmetic facts (Koponen et al.,
within RAN (digit RAN vs. nondigit RAN) and within PM (digit PM 2016; Peng et al., 2020). Articulation rates for number words are
vs. nondigit PM). We also wanted to examine whether phonological expected to strongly influence mathematic performance that
processing is uniquely related to mathematics by systematically parti- involves number counting. For example, counting strategies might
aling out the influence of nonverbal intelligence, executive function- be used during simple arithmetic problem solving (Geary, 1995;
ing, and vocabulary knowledge, using a meta-analytic structural Geary et al., 1993). Therefore, RAN should be important for
equation modeling (MASEM) analysis. Theoretically, the findings mathematics abilities among children across ages, since retrieving
PHONOLOGICAL PROCESSING SKILLS AND MATHEMATICS 3
phonological verbal codes usually contributes to different kinds of concerning the relations and manipulations between different nu-
mathematics, including number identification, counting, and the merical quantities, such as composition and decomposition of
speed of mathematics problem solving (Koponen et al., 2017). numbers. Accordingly, children learn mathematics in different
Such a retrieval role of RAN is critical in developing/learning forms, and mathematics understanding includes complicated prob-
advanced mathematics skills since retrieval can help to optimize lems of various types (Hecht et al., 2001). The Common Core
cognitive/language allocation during advanced mathematics per- State Standards Initiative provides essential components of mathe-
formance (e.g., slow retrieval can impede the efficient cognitive matics instruction across grades K–12 in the United States (National
processing during mathematics performance, Peng et al., 2020); Governors Association Center for Best Practices & Council of Chief
this is in line with many previous findings in which RAN corre- State School Officers, 2010), and researchers have conceptualized
lates with early numeracy knowledge (e.g., counting, number these standards as comprising different mathematics subskills,
knowledge) and simple arithmetic (e.g., addition and subtraction) including basic number knowledge, calculation, word problem
independently (Bull & Johnston, 1997; Chard et al., 2005). In con- solving, geometry, and algebra (Peng et al., 2016; Peng et al.,
trast, Cirino et al. (2018) found PA and RAN to be predictive of 2019). At the primary grades, the major curricular focus is con-
reading-related abilities, rather than calculation ability, when four ceptualized in three domains across cultures: understanding num-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
relatively straightforward procedurally, with such facts often being performance at around 9 years old. We anticipated that children of
directly retrieved from long-term memory; basic number knowl- younger ages and at lower grade levels would show the strongest
edge has been associated with both PA (Krajewski & Schneider, associations between phonological processing and mathematics.
2009) and RAN (Swanson & Kim, 2007). Moreover, there is a The relations of phonological processing abilities with mathe-
neural overlap between arithmetic problem solving using the fact matics between typically and atypically developing children is im-
retrieval strategy and PA in the left-temporo-parietal areas (Dehaene portant to compare, given that children with learning disabilities
et al., 2003; Grabner et al.,2009; Pugh et al., 2001; Schlaggar & might have a limited-capacity of short-term storage (Geary, 1993).
McCandliss, 2007; Vigneau et al., 2006). Thus, both basic number In addition, most struggling learners tend to use less efficient
knowledge and word problem solving appear to be associated with approaches for mathematics than direct retrieval, show weak flexi-
PA and RAN. Although children usually solve basic number knowl- bility in problem-solving, and are less accurate in performing
edge problems using fact retrieval strategies, while solving other mathematics algorithms (Zhang et al., 2014). Moreover, although
types of mathematics calculation problems with procedural strat- children with learning disabilities show deficits in the ability to
egies (Peng et al., 2017), sometimes strategy use depends partly on use retrieval-based processes (Geary et al., 2012), they rely more
the difficulty levels of the problems, on problem types, and on how on finger counting, and use more counting strategies during mathe-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
phonological processing and mathematics, such as phonological “Number” represented” number knowledge” OR “number sense”
processing components, mathematics subtypes, and sample char- OR “number line estimation; “calcul” included “calculation” OR
acteristics, have not been examined. Moreover, it is essential to “calculating.” In addition, we also searched unpublished literature
statistically control the influences of general cognitive skills and through ProQuest Dissertation and ResearchGate and contacted
general language abilities in order to further investigate the unique researchers when their relevant studies were not available. As shown
contributions of phonological processing abilities to mathematics in Figure 1, the initial database searching identified 24,400 records.
outcomes. The present study aimed to answer the following We removed duplicate articles for initial screening and ended up
research questions: with 21,514 records. Then, the authors, including one assistant pro-
fessor and two doctoral students in psychology, reviewed abstracts
1. What are the relationships between phonological process- of the articles and identified these studies for further full-text evalua-
ing abilities and mathematics? Based on previous find- tion. Afterward, 857 full-text articles were carefully assessed for eli-
ings, it was anticipated that phonological processing gibility with the following specific criteria.
abilities would be significantly correlated with mathemat- The specific criteria used to assess studies for eligibility and
ics skills. inclusion were as follows:
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
2. What are the moderators of the correlations between pho- 1. To acquire the validated effect sizes, included studies had
nological processing and mathematics? We investigated to employ quantitative tasks measuring mathematics,
how different types of phonological processing may influ- including basic number knowledge, calculation, or word
ence such relations across different types of mathematics problem solving. We excluded 127 studies for not reach-
skills. Based on previous work, we expected that PA and ing this inclusion criterion.
RAN would show stronger relations with mathematics
than would PM. Word problems were expected to be 2. Included studies had to employ quantitative tasks measuring
associated with phonological processing abilities than phonological processing skills, including PA, RAN, or PM.
would calculation and basic number knowledge. We also We excluded 201 studies for not meeting this criterion.
anticipated that their correlations would be generally
stronger among younger children than among older children. 3. Studies had to have obtained at least one simple correla-
tion coefficient between phonological processing and
3. How do relations between phonological processing abil- mathematics. We excluded 104 studies because there was
ities and mathematics change when other general skills/ no measure of effect size available.
covariates are statistically controlled? In addition to vo-
4. Studies had to focus on participants in kindergarten or
cabulary knowledge, nonverbal intelligence, and execu-
primary school, because this period represents among
tive functioning were included in the analysis model. We
the quickest developmental phases of both phonological
expected that phonological processing as indexed with
processing (Wagner et al., 1999; Wagner & Torgesen,
PA, RAN, and PM measures would still have a significant
1987) and mathematics (Fuchs et al., 2005, 2010; Geary,
relationship with each mathematics subskill even when
1993; Hecht et al., 2001) over time. We excluded 331
these cognitive skills were statistically controlled.
studies for not reaching this inclusion criterion.
Figure 1
Flow Diagram for the Search and Inclusion of Studies
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
from 90% to 99% across types of phonological processing, types of observed/true effects or outcomes could be accounted for (Kon-
mathematics, and sample characteristics criteria. stantopoulos, 2011).
Studentized residuals and Cook’s distances were used to exam-
Meta-Analytic Approach ine whether studies were outliers and/or influential in the context
of the model (Viechtbauer & Cheung, 2010). Specifically, studies
The analyses were conducted using the effect size index Pear- with a studentized residual larger than the percentile of a standard
son’s r for all outcomes. The correlations between the phonologi- normal distribution were considered potential outliers (i.e., using a
cal processing measures and the mathematics outcomes, as well
Bonferroni correction with two-sided for studies included in the
as children’s characteristics (age, grade, groups) were analyzed.
meta-analysis). Studies with a Cook’s distance larger than the me-
Cohen (1988) provided guidelines for interpreting the magnitude
dian plus six times the interquartile range of the Cook’s distances
of a correlation. Specifically, r = .10, r = .30, and r = .50 were
are considered to be influential.
recommended to be considered small, medium, and large in mag-
Furthermore, we carried out the subgroup analyses to examine
nitude, respectively. The metafor package, which provides a com-
the relation between phonological processing and mathematics
prehensive collection of functions for conducting meta-analyses in
R (Viechtbauer, 2010), was used for constructing the funnel plots with each moderator. Specifically, metaregression analyses were
and the following meta-analysis. The Egger’s test (Egger et al., conducted for the moderation effects of domain of phonological
1997; Sterne & Egger, 2005), using the standard error of the processing, types of mathematics skills, and sample characteristics
observed outcomes as the predictor, was utilized to check for fun- on the associations between phonological processing and mathe-
nel plot asymmetry. Specifically, the metafor package includes matics. Given very limited numbers of studies targeting children
functions for calculating various effect sizes for outcome meas- with special needs, various groups of children with different diffi-
ures, such as Fisher’s r-to-z-transformed correlation coefficients, culties, including those with learning disabilities, language delays,
and allows the user to fit fixed-, random-, and mixed-effects mod- or of lower intelligence, were grouped together in the atypical
els to these data. For nonindependent effect sizes or outcomes, the sample in the current analysis. Dummy coded variables were created
package also provides a function for fitting multilevel/multivariate to examine the comparisons among categorical variables (Cohen
random-effects models. The random effects robust standard error et al., 2003).
estimation technique, which was developed by Hedges et al. Finally, we conducted the analysis to examine the unique asso-
(2010), was used. Given that there existed several effect sizes per ciations between phonological processing and mathematics after
study, the rma.mv function was used to fit suitable meta-analytic controlling for general skills, using a two-stage meta-analytic struc-
multivariate/multilevel models so that the nonindependence in the tural equation modeling (metaSEM) approach (Cheung, 2014). We
PHONOLOGICAL PROCESSING SKILLS AND MATHEMATICS 7
quantified the heterogeneity across study (e.g., different measure- RAN, r = .35, 95% CI [.30, .39]; for digit RAN, r = .41, 95% CI
ment, sample, etc.) using I2 statistic (see Table 1). A random effects [.34, .47]; for nondigit RAN, r = .31, 95% CI [.26, .36]; for PM,
model was selected to account for the heterogeneity. In the first r = .28, 95% CI [.24, .31]; for digit PM, r = .27, 95% CI [.23, .32];
stage of the analysis, we used a multivariate method to combine cor- for nondigit PM, r = .28, 95% CI [.22, .33]. As shown in Table 3,
relation matrices across all studies (see Table 1). Essentially, we ran there were no significant differences between the relations of
a meta-analysis on the entire correlation matrix between all variables mathematics with digit RAN or with nondigit RAN, either for digit
of interest. On average, each cell in this correlation matrix aggre- PM or nondigit PM. PA was more strongly related to mathematics
gated 4,960 (range: 609–18,028) samples from 17 (range: 3–71) than PM, b = –.09, z = –9.22, p , .001, and RAN was more
studies. The mean sample size was 254 and the harmonic mean sam- strongly related to mathematics than PM, b = –.11, z = –9.96, p ,
ple size was 119. In the second stage, we fitted a structural equation .001. No significant difference was found across the relation of
model to the pooled correlation matrix (see Cheung & Chan, 2005; mathematics with PA and that with RAN.
Cheung & Hafdahl, 2016 for method details). We evaluated the
model fit in the second stage with the chi-square test, the confirma- Moderation Effects of Types of Mathematics Skills
tory fit index (CFI), and the root mean square error of approximation
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
(RMSEA). Models with CFI . .95, and RMSEA , .08, are consid-
ered with a good fit (Kline, 2016). ities and different mathematics skills, we finally obtained 407 cor-
relations involving basic number knowledge, 258 correlations
involving calculation, and 148 correlations involving word prob-
Results
lems. The average correlations between each of the phonological
Through influence analyses, nine effect sizes were identified as out- processing skills and the three mathematics subskills were signifi-
liers and were ruled out before further analyses were carried out. The cant: for basic number knowledge, r = .30, 95% CI [.26, .33]; for
Egger’s tests demonstrated that the funnel plots showed reasonable calculation, r = .35, 95% CI [.31, .38]; for word problem, r = .35,
symmetries in all reported correlations, suggesting that there was little 95% CI [.30, .40]. Word problem and calculation were more
influence from the publication bias (all the p values for Egger’s tests strongly related to phonological processing than basic number
were above .32). Based on inclusion criteria, we included 94 studies knowledge, b = .06/.04, z = 5.63/3.40, ps , .001, while there was
involving 135 independent samples, and 826 correlations between no difference between calculation and word problem when corre-
phonological processing and mathematics for the final analyses. Gen- lating to mathematics, b = .02, z = 1.40, p = .16.
erally, the relation between phonological processing and mathematics Concerning the relations between phonological processing and dif-
was significant, r = .33, p , .001, 95% CI [.30, .36]. In our next step, ferent mathematics indexes (accuracy and fluency), we finally
the moderation effects of domain of phonological processing, types of obtained 578 correlations involving accuracy and 241 correlations
mathematics skills, age, gender, and sample types on the relation involving fluency. The average correlations between phonological
between phonological processing and mathematics were examined. processing skills and accuracy/fluency were significant: for accuracy,
r = .33, 95% CI [.30, .37]; for calculation, r = .33, 95% CI [.29, .37].
Moderation Effects of Phonological Processing Abilities Mathematics accuracy was more strongly related to phonological
processing than mathematics fluency, b = –.02, z = –2.90, p , .01.
Regarding the relations of different phonological processing
skills to mathematics, 264 correlations involving PA, 269 correla- Moderation Effects of Sample Characteristics and
tions involving PM, and 260 correlations involving RAN were
Designs
included. As shown in Table 2, the average correlation between
mathematics and each of the three phonological processing com- The present study also focused on how sample characteristics,
ponents was significant: for PA, r = .38, 95% CI [.33, .42]; for including grade level and sample type, appear to affect the
Table 1
Correlations and Heterogeneity Statistics for 11 Included Constructs
Construct 1 2 3 4 5 6 7 8 9 10 11
1. PA — 0.744 0.622 0.685 0.782 0.303 0.696 0.853 0.624 0.747 0.568
2. RAN 0.325 — 0.665 0.709 0.738 0.537 0.745 0.667 0.275 0.736 0
3. PM 0.372 0.208 — 0.677 0.674 0.011 0.518 0.503 0.553 0.735 0.799
4. NVIQ 0.316 0.203 0.201 — 0.811 0.623 0.598 0.468 0.699 0.484 0.566
5. VOCAB 0.397 0.230 0.365 0.311 — 0 0 0.576 0.854 0.696 0.681
6. EF 0.343 0.286 0.313 0.289 0.352 — 0.609 0.383 0.672 0.701 0.559
7. NUM 0.389 0.279 0.254 0.272 0.246 0.339 — 0 0.483 0.601 0.802
8. CAL 0.400 0.347 0.282 0.338 0.326 0.332 0.456 — 0.466 0.567 0.355
9. WP 0.467 0.292 0.327 0.461 0.467 0.452 0.551 0.602 — 0.623 0.640
10. MACC 0.411 0.320 0.282 0.391 0.301 0.423 0.482 0.490 0.521 — 0.272
11. MF 0.376 0.387 0.264 0.336 0.273 0.362 0.453 0.487 0.535 0.556 —
Note. PA = phonological awareness; RAN = rapid automatized naming; PM = phonological memory; NVIQ = nonverbal IQ; VOCAB = vocabulary
knowledge; EF = executive functioning; NUM = number knowledge; CAL = calculation; WP = word problem; MACC = mathematics accuracy; MF =
mathematics fluency. Correlations are below the diagonal; heterogeneity statistics are above the diagonal. Zero values indicate homogenous correlations
(according to Q-test).
8 YANG ET AL.
Table 2
Number of Effect Sizes, Correlations, 95% Confidence Interval (CI), Between-Study Sampling
Variance (s2) of the Associations Between Phonological Processing and Mathematics
Phonological processing and mathematics
Measure K r 95% CI s2
Average correlation 819 .33 [.30, .36] .02
Types of mathematics (1)
1. Basic number knowledge 407 .30 [.26, .33] .01
2. Calculation 258 .35 [.31, .38] .02
3. Word problem 148 .35 [.30, .40] .03
Types of mathematics (2)
1. Accuracy 578 .33 [.30, .37] .02
2. Fluency 241 .33 [.29, .37] .02
Types of phonological processing
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
relations of phonological processing abilities with mathematics correlations that included kindergarteners, 292 including middle
across studies. Grades were categorized into three levels, including grades, and 81 senior grades within primary school. The average
K1 to K3 (kindergarten), Grade 1 to Grade 3 (middle grades), and correlations between phonological processing abilities and mathe-
Grade 4 to Grade 6 (senior grades). We finally obtained 197 matics among children at different grades were significant: for
Table 3
Moderation Analysis on the Associations Between Phonological Processing and Mathematics
Moderator B SE z 95% CI s2 Group comparisons
Types of mathematics (1) .02
Basic number knowledge vs. calculation .04 .01 3.40*** [.02, .06] Basic number knowledge ,
Basic number knowledge vs. word problem .06 .01 5.63*** [.04, .08] Calculation; Basic number
Calculation vs. word problem .01 .01 1.11 [.01, .03] knowledge , Word problem
Types of mathematics (2) .02
Accuracy vs. fluency .02 .01 2.90** [.04, .01] Accuracy . Fluency
Types of phonological processing .02
PA vs. RAN .02 .01 2.99** [.04, .01] PA . RAN. PM
PA vs. PM .10 .01 10.42*** [.12, .08]
RAN vs. PM .11 .01 9.79*** [.13, .09]
Sample .02
Typical vs. atypical (MD) .03 .04 .68 [.05, .10]
Grade
K1–K3 vs. G1–G3 .06 .03 1.85þ [.13, 0]
K1–K3 vs. G4–G6 .06 .03 1.88þ [.12, 0] K1-K3 . G3-G6
G1–G3 vs. G4–G6 .08 .06 1.43 [.20, .03]
Age .02 .01 3.12** [.03, .01] .02
Note. PA = phonological awareness; RAN = rapid automatized naming; PM = phonological memory; K = number of effect sizes; MD = children with
mathematics difficulties; s2 = between-study sampling variance.
þ
p = .06. ** p , .01. *** p , .001.
PHONOLOGICAL PROCESSING SKILLS AND MATHEMATICS 9
children in kindergarten, r = .36, 95% CI [.31, .41]; for children 95% CI [.15, .22]. Table 5 indicates that RAN was more strongly
from Grades 1 to 3, r = .32, 95% CI [.27, .37]; for children from associated with calculation than with basic number knowledge,
Grade 4 to Grade 6, r = .29, 95% CI [.17, .39]. Moreover, group b = .06, z = 3.87, p , .001, 95% CI [.03, .10], while it also yielded
types were categorized into typically developing children, low stronger correlations with basic number knowledge as compared
achieving children, and high achieving children. Because there has with word problems, b = –.05, z = –2.83, p , .001, 95% CI [–.08,
been a relative lack of studies among high achieving children on –.01]. Table 6 indicates that PM was more strongly associated
associations between phonological processing and mathematics, with word problems than with calculation, b = .08, z = 5.73, p ,
we only analyzed findings for the typically developing children .001, 95% CI [.05, .11], while it also reveals stronger correlations
and low achieving children. The average correlations among pho- with word problems than with basic number knowledge, b = .06,
nological processing and mathematics skills among typically z = 2.54, p , .001, 95% CI [.01, .11]. Both PA and PM were more
strongly associated with mathematics accuracy than mathematics
developing children were significant, r = .33, 95% CI [.30, .36].
fluency, b = .11/.05, z = 8.86/3.80, p , .001, 95% CI [.09, .14]/
The average correlations among phonological processing and
[.03, .08], while RAN was more strongly associated with mathe-
mathematics among atypically developing children were also sig-
matics fluency than accuracy, b = .07, z = 6.44, p , .001, 95% CI
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Table 4
Moderation Analysis on the Associations Between Phonological Awareness (PA) and Mathematics
Moderator b SE Z 95% CI s2 Group comparisons
Types of mathematics (1) .02
Basic number knowledge vs. calculation .02 .02 .82 [.02, .05] Word problem.Basic number
Basic number knowledge vs. word problem .19 .02 10.92*** [.15, .22] knowledge
Calculation vs. word problem .02 .03 .78 [.03, .08]
Types of mathematics (2) .03
Accuracy vs. fluency .11 .01 8.86*** [.14, .09] Accuracy . Fluency
Sample .03
Typical vs. atypical (MD) .05 .14 .37 [.23, .33]
Grade .03
K1–K3 vs. G1–G3 .01 .04 .26 [.07, .09]
K1–K3 vs. G4–G6 .06 .11 .59 [.15, .28]
G1–G3 vs. G4–G6 .01 .06 .23 [.13, .10]
Age .02 .01 1.66 [.04, .00] .02
Note. K = number of effect sizes; MD = children with mathematics difficulties; s = between-study sampling variance.
2
*** p , .001.
10 YANG ET AL.
Table 5
Moderation Analysis on the Associations Between Rapid Automatize Naming (RAN) and Mathematics
Moderator b SE z 95% CI s2 Group comparisons
Types of mathematics (1) .02
Basic number knowledge vs. calculation .06 .02 3.87*** [.03, .10] Word problem
Basic number knowledge vs. word problem .10 .02 5.13*** [.13, .06] , Basic number knowledge
Calculation vs. word problem .05 .02 2.83*** [.05, .10] , Calculation
Types of mathematics (2) .02
Accuracy vs. fluency .07 .01 6.44*** [.05, .10] Accuracy , Fluency
Sample .02
Typical vs. atypical (MD) .05 .05 .95 [.05, .15]
Grade .02
K1–K3 vs. G1–G3 .03 .05 .58 [.12, .07] K1-K3 . G4-G6
K1–K3 vs. G4–G6 .13 .06 2.25 [.24, .02]
G1–G3 vs. G4–G6 .06 .04 .79 [.09, .20]
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Note. K = number of effect sizes; MD = children with mathematics difficulties; s = between-study sampling variance.
2
df = 12, CFI = .992, RMSEA = .010. When mathematics fluency Results indicated that the effect sizes of the associations between
was the dependent variable, v2 = 40.6, df = 12, CFI = .985, various phonological processing subskills and mathematics were
RMSEA = .012. Due to the limited number of studies in word significant and moderate (rs = .27 .41). Moreover, the three pho-
problems (as low as three studies in a cell), we are unable to fit the nological processing abilities (i.e., PA, RAN, and PM) contributed
word problem model. Specifically, after statistically controlling differently to mathematics subskills across ages, grades, and sam-
for those cognitive abilities, including nonverbal IQ, executive ple characteristics. The relations between phonological processing
functioning, and vocabulary knowledge, we still found medium to abilities and each mathematics subskill remained medium to high
high correlations between phonological processing and various when vocabulary knowledge, nonverbal intelligence, and execu-
mathematics subskills (for number knowledge, r = .50; for calcula- tive functioning were statistically controlled.
tion, r = .35; accuracy r = .20, and for fluency r = .54). Furthermore, These positive associations were expected given previous
we fit a large MASEM model that included all the mathematics sub- research and theory: Specifically, PA is thought to help children to
skills and then constrained each parameter of interest to be equal. represent and manipulate basic math knowledge, RAN is perhaps
Unfortunately, due to the limited power and limited number of effect an indicator of quick retrieval of mathematics knowledge from
sizes across the models (see Table 7), the large model failed to con- long-term memory, and PM has been conceptualized to facilitate
verge when comparing across types of mathematics skills. working memory and reasoning during mathematics performance
and learning. Early phonological processing skills are likely to
promote math early, critical for later advanced math skills devel-
Discussion
opment (LeFevre et al., 2010; Peng et al., 2020), highlighting the
The primary goal of this meta-analysis was to determine the theorized link between early phonological processing skills and
associations of phonological processing abilities with mathematics subsequent mathematics development. Thus, although correla-
skills among children across kindergarten and primary school. tions or differences among correlations (e.g., .30 vs. .35) seem
Table 6
Moderation Analysis on the Associations Between Phonological Memory (PM) and Mathematics
Moderator b SE z 95% CI s2 Group comparisons
Types of mathematics (1) .02
Basic number knowledge vs. calculation .03 .02 1.66 [.01, .09] Basic number knowledge ,
Basic number knowledge vs. word problem .06 .02 2.54*** [.01, .11] Word problem; Calculation
Calculation vs. word problem .08 .01 5.73*** [.05, .11] , Word problem
Types of mathematics (2) .02
Accuracy vs. fluency .05 .01 3.80*** [.08, .03] Accuracy . Fluency
Sample .02
Typical vs. atypical (MD) .00 .07 .06 [.14, .13]
Grade .03
K1-K3 vs. G1-G3 .03 .05 .53 [.12, .07]
K1-K3 vs. G4-G6 .01 .08 .11 [.17, .15]
G1-G3 vs. G4-G6 .01 .04 .17 [.08, .07]
Age .00 .01 .45 [.02, .01] .02
Note. K = number of effect sizes; MD = children with mathematics difficulties; s = Between-study sampling variance.
2
*** p , .001.
PHONOLOGICAL PROCESSING SKILLS AND MATHEMATICS 11
Figure 2
The Analytic Model for Investigating the Unique Contribution of Phonological Processing
to Each Mathematics Subskill, Including Basic Number Knowledge, Calculation, and Word
Problem
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
relatively small/medium, they may still be theoretically impor- Moderation Effects of Phonological Processing Abilities
tant as a critical stepping stone for later math development.
These theoretical implications are in line with the findings from PA and RAN had robust associations in explaining mathematics
other more math-specific foundation skills (Chen & Li, 2014; abilities with effect sizes of .38 and .35, respectively. This is con-
Peng et al., 2016). For example, meta-analyses have demon- sistent with previous findings from Simmons and Singleton
strated small/medium relations between domain-specific math (2006), Göbel and Snowling (2010), and De Smedt and Boets
skills, such as number acuity, with symbolic math performance, (2010), studies with participants of different ages. The obtained
r = .20, 95% CI [.14, .26] (Chen & Li, 2014), and a small/me- significant relationship between PA and mathematics was compa-
dium relation between WM and mathematics, r = .35, 95% CI rable to one highlighted in the meta-analysis conducted by Peng et
[.32, .37] (Peng et al., 2016). The magnitudes of these correla- al. (2020; r = .35). Importantly, a number of theories have sug-
tions were similar to ours. gested that number verbal codes are consistently utilized during
mathematics tasks (e.g., Butterworth, 2005; Dehaene et al., 2003;
Table 7 LeFevre et al., 2010), and PA tasks are sensitive to underlying
Number of Studies for the Correlations of Phonological phonological representations and manipulations of these verbal
Processing Abilities and Each Mathematics Subskill mathematics tasks (Simmons & Singleton, 2008).
The obtained significant relationship between RAN and mathe-
Rapid matics was also comparable to the one reported in the meta-analy-
Phonological automatized Phonological
Mathematics subskill awareness (PA) naming (RAN) memory (PM) ses conducted by Koponen et al. (2017; r = .37) as well as Peng et
Number knowledge 37 21 23
al. (2020; r = .31). In the RAN-mathematics relationship analysis,
Calculation 44 24 22 we further distinguished and coded those empirical papers with
Word problem 19 8 8 varied naming stimuli (digit and nondigit). Results showed that
Mathematics accuracy 53 29 33 digit and nondigit RAN tasks did not significantly explain the
Mathematics fluency 33 19 19
RAN-mathematics relationship; this suggests that the possible
12 YANG ET AL.
mechanism underlying RAN-mathematics may be a domain-gen- previous meta-analysis of Koponen et al. (2017). This result is rea-
eral retrieval ability (Koponen et al., 2017), independent of mate- sonable as RAN exerts greater influence for task fluency, while
rial effects. taps accuracy to a much lesser extent. The moderator effects of
Moreover, the correlation between (digit and nondigit) PM and mathematics types can also be explained by the view that RAN
mathematics was also similar to that found in the studies by Peng reflects the rate of access to and retrieval speed of phonological
et al. (2016; r = .30) and Friso-Van den Bos et al. (2013; r = .31). representations stored in long-term memory. In contrast, mathe-
Compared with PA and RAN, PM accounted for less variability in matics accuracy showed a stronger relation with PA than did
mathematics in the present meta-analysis. There may be several mathematics fluency with PA; similar results were found in the
reasons for this slightly smaller association. First, the variance of relation of PM and mathematics subskills. This can be explained
mathematics abilities explained by PM could have been captured by the fact that empirical studies tend to assess PA and PM using
by PA and RAN. Consistent with this argument, Table 1 shows a tasks of accuracy rather than by how automatized these tasks are
substantial correlation between PM and PA (r = .37), and a moder- among young children (e.g., Alloway et al., 2005).
ate correlation between PM and RAN (r = .21). Second, most em-
pirical studies employed performance of either digit span tasks or Moderation Effects of Sample Characteristics
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
children are required to say aloud the sequence of orally presented The association of RAN with mathematics was stronger among
digits; testers record the maximum number of digits that children kindergarten children than among primary school children. This
can correctly recall (Wechsler, 1974). In word span tasks, children age-related difference in correlations might be attributable to the
need to recall lists of words of different lengths (Wagner et al., fact that rate of access to phonological information stored in long-
1999). Due to the lexical support for these familiar sound patterns term memory may be more beneficial for children in acquiring
of digits and/or words, we presume that children probably rely mathematics knowledge at earlier ages. Notably, children’s fact re-
heavily on the representation of the lexicon in phonological stor- trieval skills as indicated by RAN are less automatized among
age as a means of enlarging their memory span. A reliance on fa- younger children (Georgiou & Stewart, 2013; Yang et al., 2019),
miliar words might turn out to be a less sensitive indicator of PM but naming of RAN symbols (single digits or letters) reaches auto-
than unfamiliar spoken stimuli such as pseudo words (Baddeley et maticity earlier than most mathematics knowledge. During the
al., 1998; Köpke & Nespoulous, 2006), making PM less important process, RAN signifies the potential of an efficient system to pro-
for mathematics learning. We recommend using pseudo words cess multiple items simultaneously (Wei et al., 2015).
when assessing PM in future work, and comparing different indi- Along with grade as a significant moderator, we similarly found
cators of PM when correlating these with various mathematics that age could distinguish the relation between RAN and mathe-
abilities. matics, with children at younger ages demonstrating stronger rela-
tions. Further analyses showed that the relation of RAN with
Moderation Effects of Mathematics Types mathematics was more sensitive to the age variable than were PA
and PM with mathematics. Given that RAN clearly and explicitly
The current analysis suggests that the effect sizes of the rela- involves speed of processing, at least in part (e.g., McBride-Chang
tions among phonological processing abilities and mathematics & Kail, 2002), and that, additionally, speed of processing increases
skills were influenced by mathematics types. PA was more closely rapidly with development (e.g., Kail et al., 2013), the association
related to word problems than to basic number knowledge. This of RAN to mathematics may be particularly salient developmen-
might be because PA helps represent and manipulate the segments tally: Those who are quicker to process mathematical symbols
of a spoken language, which share more overlapping language likely perform better in a variety of mathematics tasks. RAN is at
components with word problems and is similar to the cognitive least partly an indicator of this. In contrast, tasks of PA and PM
processes of retrieving the calculation algorithm from word prob- do not typically involve an explicit speed component, so their
lems (Fuchs et al., 2014). Word problem was a stronger correlate interactions with mathematics may be less clearly influenced by
of PM than were either calculation or basic number knowledge. maturation.
This result might be due to the fact that word problems tend to be Furthermore, the relations between phonological processing
most complicated, in that children need to extract, manipulate, and abilities and mathematics did not vary between typically develop-
represent arithmetic facts when solving problems. Both working ing children and children with learning difficulties. Although chil-
memory and long-term memory are involved in solving word dren at different developmental stages use various strategies in
problems, which rely on PM as a medium. Meanwhile, RAN showed learning mathematics, these strategies may all require phonologi-
a stronger association with calculation than with word problem tasks. cal processing. For instance, using retrieval-based processes is
This is consistent with the finding from Koponen et al. (2017). These more popular among typically developing children, for whom pho-
researchers showed that RAN was a stronger correlate of mathemat- nological processing is essential (Hecht et al., 2001). In contrast,
ics when mathematics was operationalized with a calculation task counting strategies during mathematics problem solving, including
than when it was operationalized with comprehensive mathematics aloud counting, silent counting, and finger counting, are used more
achievement tests. Indeed, retrieving the digit names and operation frequently among struggling learners (Geary, 1993); phonological
symbols, as well as retrieving answers from memory, are the most processing is also highly associated with counting procedures
important processes involved in calculation; these processes arguably (Noël, 2009; Zhang et al., 2014). To be specific, an important role
tap similar capacities as those required for RAN. of PA and PM for mathematics may be in facilitating the construc-
Mathematics fluency showed a stronger relation with RAN than tion of the sound representation of mathematics terms. From this
did mathematics accuracy. This finding was consistent with the perspective, both skills should be equally important for typically
PHONOLOGICAL PROCESSING SKILLS AND MATHEMATICS 13
developing children and children with learning difficulties. In representation and also the temporary storage and efficient re-
addition, relatively few studies have involved children with special trieval of numerical knowledge.
needs as compared to typically developing children (55 vs. 637),
perhaps making it difficult to observe the subtle different relations Limitations and Future Directions
of phonological processing and mathematics across groups.
The study by Koponen et al. (2017) mainly focused on the rela- There are some aspects of these findings that should be extended
tionship between RAN and mathematics. In contrast, the present in the future. First, language and education systems across districts
meta-analysis took a step further by comparing the relations of dif- might be considered as moderators in future studies. For instance,
ferent phonological processing skills, including PA, PM, and mathematics training in Chinese, Japanese, and Korean children tend
RAN, with mathematics subskills. More importantly, we took to emerge earlier than in American children (Fuson & Kwon, 1991,
advantage of the metaSEM approach to reveal the unique contribu- 1992), and the base-10 concept in Chinese counting system facilitates
tion of each phonological processing skill to mathematics. In addition Chinese children in handling arithmetic problems with less cognitive
to the meta-analysis of zero-order correlations among phonological workload compared with children from English-speaking countries
processing and mathematics subskills, we expanded this important (Fuson & Kwon, 1991; Geary et al., 1993).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
work by conducting MASEM in order to identify which specific Second, as shown, the relation between phonological processing
This document is copyrighted by the American Psychological Association or one of its allied publishers.
mathematics subskill could be uniquely explained by phonological and mathematics highly depends upon what strategies are employed
processing. The MASEM results were generally consistent with the by children during problem solving (e.g., De Smedt et al., 2010).
above-mentioned simple meta-analysis findings. That is, after statisti- Children’s strategies used during mathematics problem solving,
cally controlling for domain-general abilities, including nonverbal including procedural strategies and fact retrieval strategies, should
IQ, executive functioning, and vocabulary knowledge, we still found be further evaluated. A supplementary interview might contribute
a small to medium correlation between phonological processing (as to analyzing children’s problem-solving procedures.
indexed with PA, RAN, and PM) and various mathematics subskills Third, although we have tried to precisely identify the variance
(i.e., number knowledge, calculation, mathematics accuracy, and in mathematics explained by vocabulary knowledge, and to sepa-
mathematics fluency), indicating the unique associations of phono- rate the contributions of vocabulary knowledge and domain-gen-
logical processing abilities in children's mathematics development eral cognitive skills (i.e., executive functioning and nonverbal IQ)
across most areas. to mathematics, ultimately, our model did not fit the data well,
However, the metaSEM conducted using word problems as the given relatively few observations, leading to a high standard error
dependent variable resulted in unsatisfactory model fit, making it in each related estimate. Thus, future studies should systematically
difficult to directly compare the partial correlation results with investigate the unique predictions of language and literacy for
simple meta-analyses. We assume that this might have been due to mathematics.
the small number of studies involving word problems compared Fourth, although we tried to group the meta-analytic SEM with
with the other mathematics subskills. Moreover, there might be age or grade additionally controlled in the correlational raw data,
multicollinearity between phonological processing and domain- we did not obtain enough data points to use either of them as a
general skills while predicting the word problems domains. For within-study covariate in the model. Lastly, we have tried to tease
similar reasons, when statistically controlling for general cognitive apart different subskills within PA (i.e., onset-rime, syllable, and
skills and vocabulary knowledge, we had tried to separate each phoneme). However, most of the studies of PA included two or
phonological processing skill in the metaSEM, but the sample size more subskills, and obtained a composite score in the correlation
was too small to run the metaSEM. We could not obtain a model tables. Although the present study tried to separate each phonolog-
fit with only a very small number of studies having all three PP ical processing ability in the metaSEM, ultimately, we failed to
measures. Furthermore, when we fit a large metaSEM model con- obtain a model fit with only a handful of studies including all three
taining all the mathematics subskills and then constrained each phonological processing measures. Future studies, hopefully with
parameter of interest to be equal. With the limited power and a more data, are needed to separate each phonological processing
limited number of effect sizes across the models, the large model skill and to investigate the unique contributions of PA, RAN, and
failed to converge when comparing across the types of mathemat- PM to each mathematics subskill, respectively. Finally, and cru-
ics skills. Such issues have been discussed by Jak and Cheung cially, the criteria for mathematics difficulties varied across stud-
(2020) that MASEM is very sensitive to missing data if the model ies, and there are a relatively small number of studies focusing on
involves only manifest measures. Nevertheless, when looking at atypically developing children. More research is needed to investi-
the trends, phonological processing (as indexed by PA, RAN, and gate the special needs of children with mathematics learning diffi-
PM) seemed to have a larger unique association with number culties during early education.
knowledge as compared with calculation, while it seemed to exert
tighter associations with mathematics fluency than that with math- Educational Implications
ematics accuracy. Our findings demonstrate that phonological
processing is important for mathematics development even with Results of this meta-analysis revealed that the strengths of the
many important variables statistically controlled. Peng et al. relations of phonological processing skills to mathematics abilities
(2020) did not statistically control for vocabulary knowledge in were influenced by the nature of the phonological processing com-
their analyses, and the current findings are particularly strong ponent, the types of mathematics skills assessed, and children’s
because we controlled for vocabulary knowledge, nonverbal own characteristics. With general vocabulary knowledge, nonver-
IQ, and executive functioning. Results are clear: Phonological bal reasoning, and executive functioning statistically controlled,
processing appears to support both the formation of numerical PA, RAN, and PM were all uniquely associated with mathematics
14 YANG ET AL.
subskills. Findings suggest that early PA and RAN abilities in partic- arithmetic. South African Journal of Childhood Education, 3(1), 1–20.
ular may be good early indicators of mathematics performance. In https://doi.org/10.4102/sajce.v3i1.29
order for children to learn to recognize numerical information, Baddeley, A., Gathercole, S., & Papagno, C. (1998). The phonological
count, and reason about mathematical associations, phonological loop as a language learning device. Psychological Review, 105(1),
processing is necessary; it offers a tool with which to predict and 158–173. https://doi.org/10.1037/0033-295X.105.1.158
*Barrouillet, P., & Lépine, R. (2005). Working memory and children’s use
intervene among children who are at risk of mathematics disabilities.
of retrieval to solve addition problems. Journal of Experimental Child
The finding of stronger relations of phonological processing abilities
Psychology, 91(3), 183–204. https://doi.org/10.1016/j.jecp.2005.03.002
and mathematics for younger children suggests that using phonolog- *Berg, D. H. (2008). Working memory and arithmetic calculation in chil-
ical principles might be more beneficial among children at earlier dren: The contributory roles of processing speed, short-term memory,
stages of schooling. Overall, findings indicate that early accuracy and reading. Journal of Experimental Child Psychology, 99(4),
and retrieval practice in mathematics may help to strengthen the nu- 288–308. https://doi.org/10.1016/j.jecp.2007.12.002
merical representation and retrieval of such representation, facilitat- Bireta, T. J., Fry, S. E., Jalbert, A., Neath, I., Surprenant, A. M., Tehan, G.,
ing early math performance, which relies heavily on fact retrieval. & Tolan, G. A. (2010). Backward recall and benchmark effects of work-
The current findings suggest that it may be beneficial to include ing memory. Memory & Cognition, 38(3), 279–291. https://doi.org/10
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
.3758/MC.38.3.279
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Research Synthesis Methods, 7(2), 112–120. https://doi.org/10.1002/ *Foster, M. E., Anthony, J. L., Clements, D. H., & Sarama, J. H. (2015).
jrsm.1212 Processes in the development of mathematics in kindergarten children
*Child, A. E., Cirino, P. T., Fletcher, J. M., Willcutt, E. G., & Fuchs, L. S. from Title 1 schools. Journal of Experimental Child Psychology, 140,
(2019). A cognitive dimensional approach to understanding shared and 56–73. https://doi.org/10.1016/j.jecp.2015.07.004
unique contributions to reading, math, and attention skills. Journal of Friso-Van den Bos, I., Van der Ven, S. H., Kroesbergen, E. H., & Van
Learning Disabilities, 52(1), 15–30. https://doi.org/10.1177/0022219 Luit, J. E. (2013). Working memory and mathematics in primary school
418775115 children: A meta-analysis. Educational Research Review, 10, 29–44.
Cho, J. R., & McBride-Chang, C. (2005). Levels of phonological aware- https://doi.org/10.1016/j.edurev.2013.05.003
ness in Korean and English: A 1-year longitudinal study. Journal of *Fuchs, L. S., Compton, D. L., Fuchs, D., Paulsen, K., Bryant, J. D., &
Educational Psychology, 97(4), 564–571. https://doi.org/10.1037/0022 Hamlett, C. L. (2005). The prevention, identification, and cognitive
-0663.97.4.564 determinants of math difficulty. Journal of Educational Psychology,
*Cirino, P. T. (2011). The interrelationships of mathematical precursors in 97(3), 493–513. https://doi.org/10.1037/0022-0663.97.3.493
kindergarten. Journal of Experimental Child Psychology, 108(4), *Fuchs, L. S., Compton, D. L., Fuchs, D., Powell, S. R., Schumacher,
713–733. https://doi.org/10.1016/j.jecp.2010.11.004 R. F., Hamlett, C. L., Vernier, E., Namkung, J. M., & Vukovic, R. K.
*Cirino, P. T., Child, A. E., & MacDonald, K. (2018). Longitudinal predic- (2012). Contributions of domain-general cognitive resources and different
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
tors of the overlap between reading and math skills. Contemporary Edu- forms of arithmetic development to pre-algebraic knowledge. Develop-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
cational Psychology, 54, 99–111. https://doi.org/10.1016/j.cedpsych mental Psychology, 48(5), 1315–1326. https://doi.org/10.1037/a0027475
.2018.06.002 Fuchs, L. S., Fuchs, D., Compton, D. L., Hamlett, C. L., & Wang, A. Y.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2015). Is word-problem solving a form of text comprehension? Scien-
(2nd ed.). Erlbaum. tific Studies of Reading, 19(3), 204–223. https://doi.org/10.1080/
Cohen, J., Cohen, P., West, S., & Aiken, L. (2003). Applied multiple 10888438.2015.1005745
regression/correlation analysis for the behavioral sciences (3rd ed.). Fuchs, L. S., Fuchs, D., Compton, D. L., Powell, S. R., Seethaler, P. M.,
Erlbaum. Capizzi, A. M., Schatschneider, C., & Fletcher, J. M. (2006). The cogni-
*Cowan, R., Donlan, C., Shepherd, D. L., Cole-Fletcher, R., Saxton, M., & tive correlates of third-grade skill in arithmetic, algorithmic computa-
Hurry, J. (2011). Basic calculation proficiency and mathematics tion, and arithmetic word problems. Journal of Educational Psychology,
achievement in elementary school children. Journal of Educational Psy- 98(1), 29–43. https://doi.org/10.1037/0022-0663.98.1.29
chology, 103(4), 786–803. https://doi.org/10.1037/a0024556 *Fuchs, L. S., Fuchs, D., Hamlett, C. L., Lambert, W., Stuebing, K., &
*Cowan, R., & Powell, D. (2014). The contributions of domain-general Fletcher, J. M. (2008). Problem solving and computational skill: Are
and numerical factors to third-grade arithmetic skills and mathematical they shared or distinct aspects of mathematical cognition? Journal of
learning disability. Journal of Educational Psychology, 106(1), 214–229. Educational Psychology, 100(1), 30–47. https://doi.org/10.1037/0022
https://doi.org/10.1037/a0034097 -0663.100.1.30
*Cui, J., Georgiou, G. K., Zhang, Y., Li, Y., Shu, H., & Zhou, X. (2017). *Fuchs, L. S., Fuchs, D., Hamlett, C. L., & Stecker, P. M. (1990). The role of
Examining the relationship between rapid automatized naming and skills analysis in curriculum-based measurement in math. School Psychology
arithmetic fluency in Chinese kindergarten children. Journal of Experi- Review, 19(1), 6–22. https://doi.org/10.1080/02796015.1990.12087335
mental Child Psychology, 154, 146–163. https://doi.org/10.1016/j.jecp *Fuchs, L. S., Geary, D. C., Compton, D. L., Fuchs, D., Hamlett, C. L.,
.2016.10.008 Seethaler, P. M., Bryant, J. D., & Schatschneider, C. (2010). Do differ-
*Cupples, L., Ching, T. Y., Crowe, K., Day, J., & Seeto, M. (2014). Pre- ent types of school mathematics development depend on different con-
dictors of early reading skill in 5-year-old children with hearing loss stellations of numerical versus general cognitive abilities? Developmental
who use spoken language. Reading Research Quarterly, 49(1), 85–104. Psychology, 46(6), 1731–1746. https://doi.org/10.1037/a0020662
https://doi.org/10.1002/rrq.60 Fuchs, L. S., Powell, S. R., Cirino, P. T., Schumacher, R. F., Marrin, S.,
de Jong, P. F., & van der Leij, A. (1999). Specific contributions of phono- Hamlett, C. L., Fuchs, D., Compton, D. L., & Changas, P. C. (2014).
logical abilities to early reading acquisition: Results from a Dutch latent Does calculation or word-problem instruction provide a stronger route
variable longitudinal study. Journal of Educational Psychology, 91(3), to prealgebraic knowledge? Journal of Educational Psychology, 106(4),
450–476. https://doi.org/10.1037/0022-0663.91.3.450 990–1006. https://doi.org/10.1037/a0036793
De Smedt, B., & Boets, B. (2010). Phonological processing and arithmetic *Fung, W. W. Y. (2015). Working memory components as predictors of
fact retrieval: Evidence from developmental dyslexia. Neuropsychologia, word problem solving: Does rapid automatized naming speed mediate
48(14), 3973–3981. https://doi.org/10.1016/j.neuropsychologia.2010.10 the relationship? [Doctoral dissertation]. University of California
.018 Riverside.
De Smedt, B., Taylor, J., Archibald, L., & Ansari, D. (2010). How is pho- Fuson, K. C., & Kwon, Y. (1991). Chinese-based regular and European
nological processing related to individual differences in children’s arith- irregular systems of number words: The disadvantages for English
metic skills? Developmental Science, 13(3), 508–520. https://doi.org/10 speaking children. In K. Durkin & B. Shire (Eds.), Language and mathe-
.1111/j.1467-7687.2009.00897.x matical education (pp. 211–216). Open University Press.
Dehaene, S. (1992). Varieties of mathematical abilities. Cognition, 67, Fuson, K. C., & Kwon, Y. (1992). Korean children’s single-digit addition
353–364. https://doi.org/10.1016/0010-0277(92)90049-N and subtraction: Numbers structured by ten. Journal for Research in
Dehaene, S., Piazza, M., Pinel, P., & Cohen, L. (2003). Three parietal cir- Mathematics Education, 23(2), 148–165. https://doi.org/10.2307/
cuits for number processing. Cognitive Neuropsychology, 20(3), 749498
487–506. https://doi.org/10.1080/02643290244000239 Gathercole, S. E., & Adams, A.-M. (1993). Phonological working memory
*Durand, M., Hulme, C., Larkin, R., & Snowling, M. (2005). The cogni- in very young children. Developmental Psychology, 29(4), 770–778.
tive foundations of reading and arithmetic skills in 7- to 10-year-olds. https://doi.org/10.1037/0012-1649.29.4.770
Journal of Experimental Child Psychology, 91(2), 113–136. https://doi Geary, D. C. (1993). Mathematical disabilities: Cognitive, neuropsycho-
.org/10.1016/j.jecp.2005.01.003 logical, and genetic components. Psychological Bulletin, 114(2),
Egger, M., Davey Smith, G., Schneider, M., & Minder, C. (1997). Bias in 345–362. https://doi.org/10.1037/0033-2909.114.2.345
meta-analysis detected by a simple, graphical test. British Medical Jour- Geary, D. C. (1995). Reflections of evolution and culture in children’s cog-
nal, 315(7109), 629–634. https://doi.org/10.1136/bmj.315.7109.629 nition. Implications for mathematical development and instruction.
16 YANG ET AL.
American Psychologist, 50(1), 24–37. https://doi.org/10.1037/0003-066 *Kail, R., & Hall, L. K. (1999). Sources of developmental change in child-
X.50.1.24 ren’s word-problem performance. Journal of Educational Psychology,
Geary, D. C., Bow-Thomas, C. C., Fan, L., & Siegler, R. S. (1993). Even 91(4), 660–668. https://doi.org/10.1037/0022-0663.91.4.660
before formal instruction, Chinese children outperform American chil- Kail, R. V., McBride-Chang, C., Ferrer, E., Cho, J. R., & Shu, H. (2013).
dren in mental addition. Cognitive Development, 8(4), 517–529. https:// Cultural differences in the development of processing speed. Develop-
doi.org/10.1016/S0885-2014(05)80007-3 mental Science, 16(3), 476–483. https://doi.org/10.1111/desc.12039
Geary, D. C., Hoard, M. K., & Bailey, D. H. (2012). Fact retrieval deficits *Kleemans, T., Segers, E., & Verhoeven, L. (2011a). Cognitive and lin-
in low achieving children and children with mathematical learning dis- guistic precursors to numeracy in kindergarten: Evidence from first and
ability. Journal of Learning Disabilities, 45(4), 291–307. https://doi.org/ second language learners. Learning and Individual Differences, 21(5),
10.1177/0022219410392046 555–561. https://doi.org/10.1016/j.lindif.2011.07.008
Georgiou, G., & Stewart, B. (2013). Is rapid automatized naming auto- *Kleemans, T., Segers, E., & Verhoeven, L. (2011b). Precursors to numer-
matic? Preschool and Primary Education, 1, 67–81. https://doi.org/10 acy in kindergartners with specific language impairment. Research in
.12681/ppej.46 Developmental Disabilities, 32(6), 2901–2908. https://doi.org/10.1016/j
*Georgiou, G. K., Tziraki, N., Manolitsis, G., & Fella, A. (2013). Is rapid .ridd.2011.05.013
*Kleemans, T., Segers, E., & Verhoeven, L. (2012). Naming speed as a
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
reason(s)? A follow-up study from kindergarten to Grade 1. Journal of clinical marker in predicting basic calculation skills in children with spe-
Experimental Child Psychology, 115(3), 481–496. https://doi.org/10 cific language impairment. Research in Developmental Disabilities,
.1016/j.jecp.2013.01.004 33(3), 882–889. https://doi.org/10.1016/j.ridd.2011.12.007
Göbel, S. M., & Snowling, M. J. (2010). Number-processing skills in *Kleemans, T., Segers, E., & Verhoeven, L. (2014). Cognitive and linguis-
adults with dyslexia. Quarterly Journal of Experimental Psychology: tic predictors of basic arithmetic skills: Evidence from first-language and
Human Experimental Psychology, 63(7), 1361–1373. https://doi.org/10 second-language learners. International Journal of Disability Develop-
.1080/17470210903359206 ment and Education, 61(3), 306–316. https://doi.org/10.1080/1034912X
.2014.934017
Grabner, R. H., Ansari, D., Koschutnig, K., Reishofer, G., Ebner, F., &
Kline, R. B. (2016). Principles and practice of structural equation model-
Neuper, C. (2009). To retrieve or to calculate? Left angular gyrus medi-
ing (4th ed.). Guilford Press.
ates the retrieval of arithmetic facts during problem solving. Neuropsy-
Konstantopoulos, S. (2011). Fixed effects and variance components esti-
chologia, 47(2), 604–608. https://doi.org/10.1016/j.neuropsychologia
.2008.10.013 mation in three-level meta-analysis. Research Synthesis Methods, 2(1),
61–76. https://doi.org/10.1002/jrsm.35
*Hannula, M. M., Lepola, J., & Lehtinen, E. (2010). Spontaneous focusing
Köpke, B., & Nespoulous, J. L. (2006). Working memory performance in
on numerosity as a domain-specific predictor of arithmetical skills. Jour-
expert and novice interpreters. Interpreting, 8(1), 1–23. https://doi.org/
nal of Experimental Child Psychology, 107(4), 394–406. https://doi.org/
10.1075/intp.8.1.02kop
10.1016/j.jecp.2010.06.004
*Koponen, T., Aunola, K., Ahonen, T., & Nurmi, J. E. (2007). Cognitive
*Hart, S. A., Petrill, S. A., Thompson, L. A., & Plomin, R. (2009). The
predictors of single-digit and procedural calculation skills and their
ABCs of math: A genetic analysis of mathematics and its links with
covariation with reading skill. Journal of Experimental Child Psychol-
reading ability and general cognitive ability. Journal of Educational
ogy, 97(3), 220–241. https://doi.org/10.1016/j.jecp.2007.03.001
Psychology, 101(2), 388–402. https://doi.org/10.1037/a0015115
*Koponen, T., Eklund, K., Heikkilä, R., Salminen, J., Fuchs, L., Fuchs, D.,
*Hecht, S. A., Torgesen, J. K., Wagner, R. K., & Rashotte, C. A. (2001).
& Aro, M. (2020). Cognitive correlates of the covariance in reading and
The relations between phonological processing abilities and emerging
arithmetic fluency: Importance of Serial retrieval fluency. Child Devel-
individual differences in mathematical computation skills: A longitudi-
opment, 91(4), 1063–1080. https://doi.org/10.1111/cdev.13287
nal study from second to fifth grades. Journal of Experimental Child
Koponen, T., Georgiou, G., Salmi, P., Leskinen, M., & Aro, M. (2017). A
Psychology, 79(2), 192–227. https://doi.org/10.1006/jecp.2000.2586
meta-analysis of the relation between RAN and mathematics. Journal of
Hedges, L. V., Tipton, E., & Johnson, M. C. (2010). Robust variance esti- Educational Psychology, 109(7), 977–992. https://doi.org/10.1037/
mation in meta-regression with dependent effect size estimates. edu0000182
Research Synthesis Methods, 1(1), 39–65. https://doi.org/10.1002/jrsm.5 *Koponen, T., Salmi, P., Eklund, K., & Aro, T. (2013). Counting and
*Holmes, J. (2005). Working memory and children's mathematical skills RAN: Predictors of arithmetic calculation and reading fluency. Journal
[Durham theses, Durham University]. Durham E-Theses Online. http:// of Educational Psychology, 105(1), 162–175. https://doi.org/10.1037/
etheses.dur.ac.uk/2205/ a0029285
*Hornung, C., Schiltz, C., Brunner, M., & Martin, R. (2014). Predicting *Koponen, T., Salmi, P., Torppa, M., Eklund, K., Aro, T., Aro, M.,
first-grade mathematics achievement: The contributions of domain-gen- Poikkeus, A.-M., Lerkkanen, M.-K., & Nurmi, J.-E. (2016). Counting
eral cognitive abilities, nonverbal number sense, and early number com- and rapid naming predict the fluency of arithmetic and reading skills.
petence. Frontiers in Psychology, 5, 272. https://doi.org/10.3389/fpsyg Contemporary Educational Psychology, 44-45, 83–94. https://doi.org/10
.2014.00272 .1016/j.cedpsych.2016.02.004
*Imbo, I., & Vandierendonck, A. (2007). The development of strategy use *Korhonen, J., Nyroos, M., Jonsson, B., & Eklöf, H. (2018). Additive and
in elementary school children: Working memory and individual differ- multiplicative effects of working memory and test anxiety on mathemat-
ences. Journal of Experimental Child Psychology, 96(4), 284–309. ics performance in grade 3 students. Educational Psychology, 38(5),
https://doi.org/10.1016/j.jecp.2006.09.001 572–595. https://doi.org/10.1080/01443410.2017.1356449
Jak, S., & Cheung, M. W.-L. (2020). Meta-analytic structural equation *Korpipää, H., Koponen, T., Aro, M., Tolvanen, A., Aunola, K., Poikkeus,
modeling with moderating effects on SEM parameters. Psychological A.-M., Lerkkanen, M.-K., & Nurmi, J.-E. (2017). Covariation between read-
Methods, 25(4), 430–455. https://doi.org/10.1037/met0000245 ing and arithmetic skills from Grade 1 to Grade 7. Contemporary Educational
*Jap, B. A. J., Borleffs, E., & Maassen, B. A. M. (2017). Towards identify- Psychology, 51, 131–140. https://doi.org/10.1016/j.cedpsych.2017.06.005
ing dyslexia in Standard Indonesian: The development of a reading *Krajewski, K., & Schneider, W. (2009). Exploring the impact of phonologi-
assessment battery. Reading and Writing, 30(8), 1729–1751. https://doi cal awareness, visual-spatial working memory, and preschool quantity-
.org/10.1007/s11145-017-9748-y number competencies on mathematics achievement in elementary school:
PHONOLOGICAL PROCESSING SKILLS AND MATHEMATICS 17
Findings from a 3-year longitudinal study. Journal of Experimental Child German orthography. Scientific Studies of Reading, 13(1), 1–25. https://
Psychology, 103(4), 516–531. https://doi.org/10.1016/j.jecp.2009.03.009 doi.org/10.1080/10888430802631684
*Kroesbergen, E. H., Van de Rijt, B. A. M., & Van Luit, J. E. H. (2007). work- *Moll, K., Snowling, M. J., Göbel, S. M., & Hulme, C. (2015). Early lan-
ing memory and early mathematics: possibilities for early identification of guage and executive skills predict variations in number and arithmetic
mathematics learning disabilities. Advances in Learning and Behavioral Dis- skills in children at family-risk of dyslexia and typically developing con-
abilities, 20, 1–19. https://doi.org/10.1016/S0735-004X(07)20001-1 trols. Learning and Instruction, 38, 53–62. https://doi.org/10.1016/j
*Kyttälä, M., Aunio, P., & Hautamäki, J. (2010). Working memory resour- .learninstruc.2015.03.004
ces in young children with mathematical difficulties. Scandinavian Jour- Morrison, F. J., Smith, L., & Dow-Ehrensberger, M. (1995). Education and
nal of Psychology, 51(1), 1–15. https://doi.org/10.1111/j.1467-9450 cognitive development: A natural experiment. Developmental Psychol-
.2009.00736.x ogy, 31(5), 789–799. https://doi.org/10.1037/0012-1649.31.5.789
*Kyttälä, M., Aunio, P., Lepola, J., & Hautamäki, J. (2014). The role of National Governors Association Center for Best Practices & Council of
the working memory and language skills in the prediction of word prob- Chief State School Officers. (2010). Common core state standards
lem solving in 4- to 7-year-old children. Educational Psychology, 34(6), mathematics. National Governors Association Center for Best Practices
674–696. https://doi.org/10.1080/01443410.2013.814192 & Council of Chief State School Officers.
LeFevre, J. A., Fast, L., Skwarchuk, S. L., Smith-Chant, B. L., Bisanz, J., *Navarro, J. I., Aguilar, M., Alcalde, C., Ruiz, G., Marchena, E., &
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Kamawar, D., & Penner-Wilger, M. (2010). Pathways to mathematics: Menacho, I. (2011). Inhibitory processes, working memory, phonologi-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Longitudinal predictors of performance. Child Development, 81(6), cal awareness, naming speed, and early arithmetic achievement. The
1753–1767. https://doi.org/10.1111/j.1467-8624.2010.01508.x Spanish Journal of Psychology, 14(2), 580–588. https://doi.org/10.5209/
Li, S. C., & Lewandowsky, S. (1995). Forward and backward recall: Dif- rev_SJOP.2011.v14.n2.6
ferent retrieval processes. Journal of Experimental Psychology: Learn- *Noël, M. P. (2009). Counting on working memory when learning to count
ing, Memory, and Cognition, 21, 837–847. https://doi.org/10.1037/0278 and to add: A preschool study. Developmental Psychology, 45(6),
-7393.21.4.837 1630–1643. https://doi.org/10.1037/a0016224
Liberman, I. Y., Shankweiler, D., Fischer, F. W., & Carter, B. (1974). Noel, M., Seron, X., & Trovarelli, F. (2004). Working memory as a predic-
Explicit syllable and phoneme segmentation in the young child. Journal tor of addition skills and addition strategies in children. Cahiers de Psy-
of Experimental Child Psychology, 18(2), 201–212. https://doi.org/10 chologie Cognitive, 22, 3–25.
.1016/0022-0965(74)90101-5 OECD. (2014). PISA 2012 results: What students know and can do—
Mann, V. A., & Liberman, I. Y. (1984). Phonological awareness and Student performance in mathematics, reading and science (Vol. I).
PISA, OECD Publishing. https://doi.org/10.1787/9789264201118-en
verbal short-term memory. Journal of Learning Disabilities, 17(10),
Onochie-Quintanilla, E., Defior, S. A., & Simpson, I. C. (2019). RAN and
592–599. https://doi.org/10.1177/002221948401701005
orthographic processing: What can syllable frequency tell us about this
*Manolitsis, G., Georgiou, G. K., & Tziraki, N. (2013). Examining the
relationship? Journal of Experimental Child Psychology, 182, 1–17.
effects of home literacy and numeracy environment on early reading and
https://doi.org/10.1016/j.jecp.2019.01.002
math acquisition. Early Childhood Research Quarterly, 28(4), 692–703.
*Östergren, R., & Träff, U. (2013). Early number knowledge and cognitive
https://doi.org/10.1016/j.ecresq.2013.05.004
ability affect early arithmetic ability. Journal of Experimental Child Psy-
*Martin, R. B. (2011). Counting procedural skill and conceptual knowl-
chology, 115(3), 405–421. https://doi.org/10.1016/j.jecp.2013.03.007
edge in kindergarten as predictors of Grade 1 math skills [Doctoral dis-
Paas, F., van Gog, T., & Sweller, J. (2010). Cognitive load theory: New
sertation]. University of Houston.
conceptualizations, specifications, and integrated research perspectives.
*Mazzocco, M. M. M., & Myers, G. F. (2003). Complexities in identifying
Educational Psychology Review, 22(2), 115–121. https://doi.org/10.100
and defining mathematics learning disability in the primary school-age
7/s10648-010-9133-8
years. Annals of Dyslexia, 53(1), 218–253. https://doi.org/10.1007/ *Passolunghi, M. C., & Lanfranchi, S. (2012). Domain-specific and do-
s11881-003-0011-7
main-general precursors of mathematical achievement: A longitudinal
McBride, C. (2016). Children’s literacy development: A cross-cultural study from kindergarten to first grade. The British Journal of Educa-
perspective on learning to read and write (2nd ed.). Routledge. tional Psychology, 82(Pt. 1), 42–63. https://doi.org/10.1111/j.2044-8279
McBride-Chang, C., & Ho, C. S. H. (2000). Developmental issues in Chi- .2011.02039.x
nese children’s character acquisition. Journal of Educational Psychol- *Passolunghi, M. C., & Siegel, L. S. (2004). Working memory and access
ogy, 92(1), 50–61. https://doi.org/10.1037/0022-0663.92.1.50 to numerical information in children with disability in mathematics.
McBride-Chang, C., & Kail, R. V. (2002). Cross-cultural similarities in the Journal of Experimental Child Psychology, 88(4), 348–367. https://doi
predictors of reading acquisition. Child Development, 73(5), 1392–1407. .org/10.1016/j.jecp.2004.04.002
https://doi.org/10.1111/1467-8624.00479 *Passolunghi, M. C., Lanfranchi, S., Altoè, G., & Sollazzo, N. (2015).
*McGrew, K. S., & Hessler, G. L. (1995). The relationship between the Early numerical abilities and cognitive skills in kindergarten children.
WJ-R Gf-Gc cognitive clusters and mathematics achievement across the Journal of Experimental Child Psychology, 135, 25–42. https://doi.org/
life-span. Journal of Psychoeducational Assessment, 13(1), 21–38. 10.1016/j.jecp.2015.02.001
https://doi.org/10.1177/073428299501300102 *Passolunghi, M. C., Mammarella, I. C., & Altoè, G. (2008). Cognitive
*Michalczyk, K., Krajewski, K., Prebler, A. L., & Hasselhorn, M. (2013). abilities as precursors of the early acquisition of mathematical skills dur-
The relationships between quantity-number competencies, working ing first through second grades. Developmental Neuropsychology, 33(3),
memory, and phonological awareness in 5- and 6-year-olds. British 229–250. https://doi.org/10.1080/87565640801982320
Journal of Developmental Psychology, 31(4), 408–424. https://doi.org/ *Passolunghi, M. C., Vercelloni, B., & Schadee, H. (2007). The precursors
10.1111/bjdp.12016 of mathematics learning: Working memory, phonological ability and nu-
Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred merical competence. Cognitive Development, 22(2), 165–184. https://
reporting items for systematic reviews and meta-analyses: The PRISMA doi.org/10.1016/j.cogdev.2006.09.001
statement. Annals of Internal Medicine, 151(4), 264–269. https://doi *Pauly, H., Linkersdörfer, J., Lindberg, S., Woerner, W., Hasselhorn, M.,
.org/10.7326/0003-4819-151-4-200908180-00135 & Lonnemann, J. (2011). Domain-specific Rapid Automatized Naming
Moll, K., Fussenegger, B., Willburger, E., & Landerl, K. (2009). RAN is deficits in children at risk for learning disabilities. Journal of Neurolinguistics,
not a measure of orthographic processing: Evidence from the asymmetric 24(5), 602–610. https://doi.org/10.1016/j.jneuroling.2011.02.002
18 YANG ET AL.
Pelletier, P. M., Ahmad, S. A., & Rourke, B. P. (2001). Classification rules St. Clair-Thompson, H. L. (2010). Backwards digit recall: A measure of
for basic phonological processing disabilities and nonverbal learning short-term memory of working memory? The European Journal of Cog-
disabilities: Formulation and external validity. Child Neuropsychology, nitive Psychology, 22(2), 286–297. https://doi.org/10.1080/09541440
7(2), 84–98. https://doi.org/10.1076/chin.7.2.84.3127 902771299
Peng, P., Lin, X., Ünal, Z. E., Lee, K., Namkung, J., Chow, J., & Sales, A. Sterne, J. A. C., & Egger, M. (2005). Regression methods to detect publi-
(2020). Examining the mutual relations between language and mathe- cation and other bias in meta-analysis. In H. R. Rothstein, A. J. Sutton,
matics: A meta-analysis. Psychological Bulletin, 146(7), 595–634. & M. Borenstein (Eds.), Publication bias in meta-analysis: Prevention,
https://doi.org/10.1037/bul0000231 assessment and adjustment (pp. 99–110). Wiley. https://doi.org/10
Peng, P., Namkung, J., Barnes, M., & Sun, C. (2016). A meta-analysis of .1002/0470870168.ch6
mathematics and working memory: Moderating effects of working *Swanson, H. L. (2006a). Cognitive processes that underlie mathematical
memory domain, type of mathematics skill, and sample characteristics. precociousness in young children. Journal of Experimental Child Psy-
Journal of Educational Psychology, 108(4), 455–473. https://doi.org/10 chology, 93(3), 239–264. https://doi.org/10.1016/j.jecp.2005.09.006
.1037/edu0000079 *Swanson, H. L. (2006b). Cross-sectional and incremental changes in
Peng, P., Wang, T., Wang, C., & Lin, X. (2019). A meta-analysis on the working memory and mathematical problem solving. Journal of Educa-
relation between fluid intelligence and reading/mathematics: Effects of tional Psychology, 98(2), 265–281. https://doi.org/10.1037/0022-0663
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
tasks, age, and social economics status. Psychological Bulletin, 145(2), .98.2.265
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Vigneau, M., Beaucousin, V., Hervé, P. Y., Duffau, H., Crivello, F., Yang, X., & McBride, C. (2020). How do phonological processing abilities
Houdé, O., Mazoyer, B., & Tzourio-Mazoyer, N. (2006). Meta-analyz- contribute to early Chinese reading and mathematics? Educational Psy-
ing left hemisphere language areas: Phonology, semantics, and sentence chology, 40, 893–911. https://doi.org/10.1080/01443410.2020.1771679
processing. NeuroImage, 30(4), 1414–1432. https://doi.org/10.1016/j Yang, X., McBride, C., Ho, C. S. H., & Chung, K. K. H. (2019). Longitu-
.neuroimage.2005.11.002 dinal associations of phonological processing skills, Chinese word read-
Wagner, R. K. (1986). Phonological processing abilities and reading: ing, and arithmetic. Reading and Writing, 33, 1679–1699. https://doi
Implications for disabled readers. Journal of Learning Disabilities, .org/10.1007/s11145-019-09998-9
19(10), 623–630. https://doi.org/10.1177/002221948601901009 Zhang, D., Ding, Y., Barrett, D. E., Xin, Y. P., & Liu, R. (2014). A com-
Wagner, R. K., & Torgesen, J. K. (1987). The nature of phonological proc- parison of strategic development for multiplication problem solving in
essing and its causal role in the acquisition of reading skills. Psychologi- low-, average-, and high-achieving students. European Journal of Psy-
cal Bulletin, 101(2), 192–212. https://doi.org/10.1037/0033-2909.101.2 chology of Education, 29(2), 195–214. https://doi.org/10.1007/s10212
.192 -013-0194-1
Wagner, R. K., Torgesen, J. K., & Rashotte, A. C. (1994). Development of *Zhang, X., Koponen, T., Räsänen, P., Aunola, K., Lerkkanen, M. K., &
reading-related phonological processing abilities: New evidence of Nurmi, J. E. (2014). Linguistic and spatial skills predict early arithmetic
bidirectional causality from a latent variable longitudinal study. Devel- development via counting sequence knowledge. Child Development,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
.30.1.73 *Zhang, X., Räsänen, P., Koponen, T., Aunola, K., Lerkkanen, M. K., &
Wagner, R. K., Torgesen, J. K., Rashotte, A. C., & Pearson, N. A. (1999). Nurmi, J. E. (2017). Knowing, applying, and reasoning about arithmetic:
Comprehensive test of phonological processing: CTOPP. Pro-ed. Roles of domain-general and numerical skills in multiple domains of
Wagner, R., Balthazor, M., Hurley, S., Morgan, S., Rashotte, C., Shaner, arithmetic learning. Developmental Psychology, 53(12), 2304–2318.
R., Simmons, K., & Stage, S. (1987). The nature of prereaders’ phono- https://doi.org/10.1037/dev0000432
logical processing abilities. Cognitive Development, 2(4), 355–373. *Zheng, X., Swanson, H. L., & Marcoulides, G. A. (2011). Working mem-
https://doi.org/10.1016/S0885-2014(87)80013-8 ory components as predictors of children’s mathematical word problem
Wechsler, D. (1974). Manual for the Wechsler intelligence scale for chil- solving. Journal of Experimental Child Psychology, 110(4), 481–498.
dren, revised. Psychological Corporation. https://doi.org/10.1016/j.jecp.2011.06.001
Wei, W., Georgiou, G. K., & Deng, C. (2015). Examining the cross-lagged *Zhou, X., Wei, W., Zhang, Y., Cui, J., & Chen, C. (2015). Visual percep-
relationships between RAN and word reading in Chinese. Scientific tion can account for the close relation between numerosity processing
Studies of Reading, 19(6), 446–455. https://doi.org/10.1080/10888438 and computational fluency. Frontiers in Psychology, 6, 1364. https://doi
.2015.1077447 .org/10.3389/fpsyg.2015.01364
Wilkinson, G. S., & Robertson, G. J. (2006). Wide range achievement test
(WRAT4). Psychological Assessment Resources. Received June 26, 2020
Woodcock, R. W., McGrew, K. S., & Mather, N. (2001). Woodcock-John- Revision received May 25, 2021
son III Tests of Achievement. Riverside Publishing. Accepted August 10, 2021 n