Cognitive Development 62 (2022) 101149

Contents lists available at ScienceDirect

Cognitive Development
journal homepage: www.elsevier.com/locate/cogdev

Bidirectional associations among executive functions,

visual-spatial skills, and mathematical achievement in primary
school students: Insights from a longitudinal study
Tobias Kahl a, *, Robin Segerer a, Alexander Grob a, Wenke Möhring a, b
a
University of Basel, Department of Psychology, Missionstrasse 62, 4055 Basel,Switzerland
b
University of Education Schwäbisch Gmünd, Department of Educational and Health Psychology, Oberbettringer Strasse 200, 73525 Schwäbisch
Gmünd, Germany

A R T I C L E I N F O A B S T R A C T

Keywords: The current study examined the dynamic and complex interplay among executive functions,
Bidirectional relations visual-spatial skills, and mathematical achievement in primary school students. In particular, we
Executive functions investigated the bidirectional, longitudinal associations among these abilities, with executive
Visual-spatial skills
functions consisting of working memory, inhibition, and cognitive flexibility. We examined 6- to
Mathematical achievement
Primary school
7-year-olds (N = 182) at the beginning of primary school and again 3 years later using the In­
telligence and Development Scales-2. A cross-lagged panel model identified asymmetrical pat­
terns in the data. Early visual-spatial skills predicted children’s later mathematical achievement;
however, working memory predictions of later mathematical achievement depended on the time
differences between assessments. Children with higher initial working memory skills showed
accelerated gains in mathematical performance when the time difference between assessments
was large. Importantly, early mathematical achievement predicted changes in later working
memory, pointing to a bidirectional relation between working memory and mathematical
achievement.

1. Introduction

Mathematical knowledge is an important pillar for children’s academic development, and influences later employability, mental
health, and life satisfaction even after accounting for reading ability and intelligence (Lyons & Huebner, 2016; Paglin & Rufolo, 1990;
Rivera-Batiz, 1992). The important role of numerical abilities and probabilistic thinking in modern societies has led to an increased
awareness in investigating the development of mathematical skills and their predicting variables. Research indicates that children
seem predisposed to numerical processing (Wynn, 1992) and that domain-specific skills such as a basic numerical understanding (e.g.,
a number sense, Dehaene, 1992) are linked to mathematical outcomes (Butterworth, 2005). Furthermore, general cognitive processes
such as processing speed, working memory or visual-spatial skills seem crucial in developing mathematical skills (Chu, van Marle, &
Geary, 2016). More concretely, it has been found that executive functions (EF) and visual-spatial skills are strongly associated with
children’s mathematical achievement (for reviews, see Cragg & Gilmore, 2014; Mix & Cheng, 2012; Newcombe, Möhring, & Frick,
2018). So far, relations among EF, visual-spatial skills, and mathematical understanding have been predominantly investigated using

* Corresponding author.
E-mail address: [email protected] (T. Kahl).

https://doi.org/10.1016/j.cogdev.2021.101149
Received 28 January 2021; Received in revised form 17 December 2021; Accepted 20 December 2021
Available online 13 January 2022
0885-2014/© 2021 Published by Elsevier Inc.
T. Kahl et al. Cognitive Development 62 (2022) 101149

cross-sectional samples indicating concurrent relations among these skills (e.g., Blair & Razza, 2007; Cragg, Keeble, Richardson,
Roome, & Gilmore, 2017; Lee et al. 2012; Monette, Bigras, & Guay, 2011). However, a growing number of longitudinal studies shows
that EF and visual-spatial skills predict later mathematical performance (e.g., Gunderson, Ramirez, Beilock, & Levine, 2012; Lee &
Bull, 2016).
Whereas these studies have improved our understanding of associations among those variables, most of these studies have
measured EF and visual-spatial skills as foundational abilities that help develop later mathematical thinking. This idea reflects a
conventional view about cognitive abilities being the basis for emerging scholastic skills (e.g., Cattell, 1987; Sternberg, Kaufman, &
Grigorenko, 2008; for a discussion, see Peng & Kievit, 2020). As a consequence, the results of these studies suggest a unidirectional
association with early EF and visual-spatial skills predicting later mathematical performance. However, relations between
domain-general abilities and mathematical skills may be more mutually connected. Several mathematical operations require working
memory, as when a person keeps an intermediate result in mind while solving an arithmetical problem. Thus, as children solve many
mathematical problems, they improve their working memory skills at the same time – even though this might be restricted to using
working memory efficiently in the mathematical domain, without necessarily improving working memory in every domain (Doebel,
2020; Peng & Goodrich, 2020). In turn, an increased working memory capacity facilitates holding mathematical information in mind,
which is again helpful for solving subsequent arithmetical problems. This example indicates that relations among cognitive and ac­
ademic variables may be more dynamic and complex than is reflected in most current research. In line with this idea, there is an
emerging interest on reciprocal relations between cognitive abilities and academic achievement (van der Maas et al. 2006; for a review,
Peng & Kievit, 2020). Such reciprocal relations might be especially prominent when children enter formal education and children’s
academic skills develop considerably. Thus, this stage might be the most important and sensitive time for bidirectionality (Peng &
Kievit, 2020); however, empirical evidence focusing on this particular stage is still scarce.

1.1. Uni- and bidirectional relations between executive functions and mathematical achievement

EF are a set of regulation and control processes that enable goal-oriented actions and management of one’s thoughts and behaviors
(Miyake & Friedman, 2012). Typically, EF are distinguished into three components: working memory, inhibition, and cognitive
flexibility (e.g., Diamond, 2013). Working memory is a cognitive store with a limited capacity that holds information in mind and the
ability to manipulate this information (Baddeley, 1996), inhibition refers to the ability to control attention or behavior to suppress a
dominant response (Diamond, 2013), and cognitive flexibility is the ability to change perspectives (Diamond, 2013). A large body of
research has shown that all components relate to mathematical achievement in various age groups across childhood and adolescence
(general overview: Bull & Scerif, 2001; working memory meta-analysis: Friso-Van den Bos et al., 2013; inhibition meta-analysis: Allan,
Hume, Allan, Farrington, & Lonigan, 2014; cognitive flexibility meta-analysis: Yeniad, Malda, Mesman, van IJzendoorn, & Pieper,
2013), with in particular, working memory being strongly related to mathematical achievement (working memory review: Raghubar
et al., 2010). These subcomponents may be related to mathematical achievement because they a) help keeping relevant information
such as intermediate results in mind (working memory), b) help suppressing automatic number facts or even procedures while solving
mathematical problems (inhibition), and c) help changing strategies or tasks when required (cognitive flexibility).
Longitudinal studies have shown that both a unified construct of EF and its individual components predict mathematical
achievement at a later point (Clark, Pritchard, & Woodward, 2010; Geary, 2011; Lee & Bull, 2016; Vandenbroucke, Verschueren, &
Baeyens, 2017). However, the vast majority did not investigate the reverse effect. That is, whether earlier mathematical achievement
may predict later EF. This lack of knowledge is predominantly based on the idea that academic achievement must have a foundation in
basic cognitive skills (Cattell, 1987), which is evident, for example, in interventions that focus on EF to improve mathematical
achievement (e.g., Dunning, Holmes, & Gathercole, 2013; Kubesch, 2016). However, there is increasing evidence that speaks for
bidirectional associations between cognitive and mathematical skills (for a review, Peng & Kievit, 2020). This bidirectionality is
explained by the transactional model (Peng & Kievit, 2020), which states that academic tasks such as mathematical skills require
cognitive abilities and practicing those academic tasks may improve cognitive abilities.
Bidirectional relations can be examined by looking at how these skills change over time, typically by using cross-lagged models.
Such an approach requires an assessment of every skill at each measurement time point in a longitudinal examination (Selig & Little,
2012). However, only few studies have implemented this approach to investigate mutuality among EF, visual-spatial skills, and
mathematical achievement (for exceptions, Feldon & Litson, 2021; Fuhs, Nesbitt, Farran, & Dong, 2014; Fung, Chung, & Lam, 2020;
Geer, Quinn, & Ganley, 2019; Miller-Cotto & Byrnes, 2019; Schmitt, Geldhof, Purpura, Duncan, & McClelland, 2017).
Overall, these latter studies have shown that core components of EF predict later mathematical achievement while at the same time,
early mathematical achievement predicts later EF. One of these studies has found that particularly working memory and mathematical
achievement predict each other from the beginning of kindergarten through second grade (Miller-Cotto & Byrnes, 2019; for an
attempted replication, see Feldon & Litson, 2021). Results of another study have indicated bidirectional relations among every core
component of EF (working memory, inhibition, and cognitive flexibility) and mathematical achievement as well as a significant
relation between children’s rate of growth in EF components and their growth in mathematical achievement over the preschool period
(Schmitt et al., 2017). In a sample of 4-year-olds, Fuhs and colleagues (2014) examined bidirectional relations between EF and
mathematical achievement across the kindergarten years. Results from this study show that EF and mathematical achievement are in
fact mutually associated across kindergarten, even though EF predicted mathematical achievement to a stronger extent than vice versa.
However, these results are contrasted by Fung et al. (2020) who investigated the bidirectional associations among EF, visual-spatial
skills, and mathematical achievement in Chinese kindergarten children in their transition to first grade. They found no relations
between early EF and later mathematical achievement nor vice versa. In sum, the majority of studies has revealed that bidirectional

2
T. Kahl et al. Cognitive Development 62 (2022) 101149

relations between EF and mathematical achievement are likely––even though this bidirectionality may change when children transfer
to primary school.

1.2. Uni- and bidirectional relations between visual-spatial skills and mathematical achievement

Visual-spatial skills are defined as the ability to create, retrieve, and transform visual images (Casey, Nuttall, & Benbow, 1995).
Visual-spatial skills comprise different subcomponents, but there is not yet a typology that the scientific community has agreed upon
(cf. Linn & Petersen, 1985; Newcombe & Shipley, 2015). Even though there is no consensus yet on the factorial structure, it has been
repeatedly shown that several types of visual-spatial skills, such as mental rotation, spatial visualization, or spatial perception
(Newcombe & Shipley, 2015; Uttal et al., 2013) relate to mathematical achievement. Mental rotation refers to recognizing rotated
objects in 2D or 3D; spatial visualization is described as the ability to assemble different pieces of objects into a more complex
configuration (Newcombe & Shipley, 2015), and spatial perception is seen as the ability to understand the spatial relationships among
objects in visual images (Uttal et al., 2013). Each of these types of visual-spatial skills is associated with mathematical achievement in
various studies (e.g., Kahl, Grob, Segerer, & Möhring, 2021; Mix & Cheng, 2012; Möhring, Frick, & Newcombe, 2018; for a
meta-analysis, Xie, Zhang, Chen, & Xin, 2020). Research on underlying mechanisms is still in its beginnings and several mechanisms
are discussed. For example, visual-spatial skills might play a key role in representing magnitude along a mental number line (Gun­
derson et al., 2012). Furthermore, they might be helpful to mentally visualize mathematical problems (Mix et al., 2016; Uttal & Cohen,
2012). Finally, visual-spatial skills might be specifically related to certain mathematical domains such as geometry.
In analogy to studies investigating associations between EF and mathematics, the vast majority of studies on visual-spatial skills and
mathematical performance has investigated relations concurrently or in a longitudinal, unidirectional manner (e.g., Frick, 2019;
Gunderson et al., 2012; Verdine, Irwin, Golinkoff, & Hirsh-Pasek, 2014). To our knowledge, there are only two studies investigating
bidirectional associations between visual-spatial skills and mathematical achievement so far. In one study, Fung and colleagues (2020)
studied visual-spatial skills in Chinese kindergarten children and found unidirectional patterns between early mental rotation, spatial
disembedding, and later math, but not vice versa. In the other study, Geer and colleagues (2019) investigated students’ spatial
visualization, spatial perception, and mathematical skills during primary school in first through third graders. Contrary to findings
from Fung and colleagues (2020), the authors showed bidirectional relations with earlier spatial visualization predicting later math
skills as well as early math skills predicting later spatial visualization skills. Importantly, no bidirectional relations were found for
spatial perception. Therefore, the few studies on this topic have revealed mixed findings so far and more studies are clearly needed.

1.3. The present study

As of today, most studies have investigated bidirectional relations between domain-general skills and mathematical achievement in
kindergarten or preschool children (Fuhs et al., 2014; Fung et al., 2020; Miller-Cotto & Byrnes, 2019; Schmitt et al., 2017). Given that
the time in primary school is highly important for developing (later) mathematical and cognitive skills (Duncan et al., 2007; Geary,
2000), studies in this age range are urgently needed. Moreover, the few studies on bidirectionality have typically examined only one of
the two domain-general skills with mathematical achievement.
Thus, the aims of the present study are twofold: First, we aim to examine the bidirectional associations among both domain-general
skills (EF and visual-spatial skills) and mathematical achievement while accounting for the respective other skill. This is important as
visual-spatial skills require executive resources and many studies have refrained to investigate both skills simultaneously. Additionally,
we controlled for maternal education, sex, and verbal reasoning based on research indicating their impact on mathematical
achievement (e.g., Geary, 2006; Hoff, 2013; Lindberg, Hyde, Petersen, & Linn, 2010; for meta-analyses, Kuncel, Hezlett, & Ones, 2004;
Peng, Wang, Wang, & Lin, 2019). Second, we aimed to investigate these relations during a period when students are continuously
confronted with formal mathematics in primary school.
The current study expands a previous study from our research group that investigated associations among EF, visual-spatial skills,
and mathematical achievement in a large sample aged 5 to 20 years (Kahl et al., 2021). Using a cross-sectional design, we were able to
show that EF and visual-spatial skills independently explained variance in participants’ mathematical achievement. However, the
cross-sectional design did not allow us to assess longitudinal, bidirectional relations among the variables of interest. Therefore, in the
current study, we aim to extend this earlier study by examining the bidirectional effects among EF, visual-spatial skills, and children’s
mathematical achievement in primary-school children.

2. Method

2.1. Participants and study procedure

Participants in the current study were recruited from a pool of children who took part in the standardization of the Intelligence and
Development Scales-2 test battery (IDS-2; Grob & Hagmann-von Arx, 2018), which we considered to be the first assessment time point
(T1). The full standardization sample included 2,030 children, adolescents, and young adults aged 5 to 20 years, examined in
Switzerland, Germany, Austria, and Liechtenstein. After an average of 3 years, we invited a subsample of this standardization sample to
participate in the present study, which we considered to be the second assessment time point (T2). Participants were re-assessed using
selected subtests on EF, visual-spatial skills, and mathematical achievement as well as another standardized test on mathematics. This
longitudinal study included children living in Switzerland who were 6 to 7 years old (6;00 - 7;11) at the time of their first assessment.

3
T. Kahl et al. Cognitive Development 62 (2022) 101149

Of the 182 participants who met these criteria, 78 participated in the second assessment (48.7% female). The assessments were
conducted by trained experimenters and took place in individual, face-to-face settings. Drop-out analyses showed that children from
families who agreed for another assessment demonstrated significantly higher scores on verbal reasoning, F(1, 180) = 14.69, p <
0.001, η2 = 0.075, working memory, F(1, 180) = 19.76, p < 0.001, η2 = 0.099, cognitive flexibility, F(1, 180) = 4.17, p < 0.05, η2 =
0.023, inhibition (F(1, 180) = 4.15, p < 0.05, η2 = 0.023, and mathematical achievement, F(1, 180) = 197.64, p < 0.001, η2 = 0.11, as
compared to children from families who did not agree to participate.
The time between T1 and T2 was 3 years on average but differed for individual children (M = 3.09, SD = 0.49, diffmin = 2.35,
diffmax = 4.28). Due to the comprehensive and time-consuming individual test sessions, the test period of the original IDS-2 stan­
dardization study (T1) extended over a period of 2 years (2015-2017). The assessments at T2 were carried out over a period of eight
months (June 2019 - February 2020).
A total of 74.3% of the subsample were monolingual native German speakers, 17.6% reported being balanced bilingual, and 8.1%
reported another dominant language. Care was taken that every participant showed a German language level that allowed them to
understand and follow the task instructions. Assessments took place individually at participants’ homes or at the laboratory of the
respective University. Parents and children provided written and oral consent for participation. Participation was compensated with a
voucher for toys up to the value of 25 Swiss franc (approximately 28 US-Dollars). Furthermore, parents received feedback on their
children’s mathematical performance. The Ethics Committee of Northwestern and Central Switzerland (EKNZ) approved the current
study which was performed in accordance with the rules laid down in the 1964 Declaration of Helsinki and its later amendments.

2.2. Measures

Variables of interest were assessed with two standardized tests at both assessment time points (i.e., IDS-2 and German Test of
Mathematical Abilities). The IDS-2 is a test battery that covers a wide range of core competencies in development and intelligence (e.g.,
EF, psychomotor skills, social-emotional competencies). Selected scales of this broad instrument were used for the present study. As is
customary in intelligence tests covering wide age ranges and proficiency levels (e.g., Wechsler Intelligence Scale for Children, WISC-IV,
Wechsler, 2003), participants started at different items of the scales in most subtests, depending on their ages. Furthermore, they
terminated the test session on varying sub-test levels depending on their performance. The variable mathematical achievement was
additionally measured with another standardized test (The German Test of Mathematical Abilities; DEMAT 2+: Krajewski, Liehm, &
Schneider, 2004; DEMAT 3+: Roick, Gölitz, & Hasselhorn, 2004; DEMAT 4: Gölitz, Roick, & Hasselhorn, 2006; DEMAT 5+: Götz,
Lingel, & Schneider, 2013) at T2 to ensure the generalizability of the findings (for details, see below). The entire test session took
approximately 2.5 hours, including breaks. For our analyses, we used age-standardized values to reduce age-specific variance and to
account for the different intervals between the assessment time points in our sample.

2.3. Mathematical achievement (IDS-2)

This subtest assessed mathematical achievement based on the Swiss school curriculum (Erziehungsdirektoren-Konferenz D-EDK,
2013). In total, this task comprised 19 items, each consisting of different sub-items. Items covered different mathematical topics with
items increasing in difficulty. Younger children (5 to 7 years) started with mathematical precursor skills (e.g., counting or knowledge
about invariance), followed by items on equations, proportions, or mental addition for children aged 8 to 10 years. Older participants
(11 years and older) started with items on proportional reasoning, geometrics, algebra, or fractions. The task was stopped when
participants answered incorrectly in five subsequent items. The first eight items were untimed whereas the following items had a time
limit of 90 seconds. The total number of correct answers served as a mathematical achievement score. The authors reported high
reliability scores in different age groups (Cronbach’s αs:.88-.92). External validity of the scale was demonstrated by significant cor­
relations with students’ math grades (r =.37, N = 478) and with parental reports on their children’s mathematical competencies (r
=.44, N = 726; Grob & Hagmann-von Arx, 2018).

2.4. Mathematical achievement (DEMAT)

The German Test of Mathematical Abilities (DEMAT) comprised a set of different standardized tests to measure children’s
mathematical skills in various grades. In accordance with the IDS-2, this test is also based on the school curriculum. As participants of
our study were in different grades at T2 (2nd to 6th grade), we used the respective versions of the DEMAT to measure their mathematical
achievement (DEMAT 2+: Krajewski et al., 2004; DEMAT 3+: Roick et al., 2004; DEMAT 4: Gölitz et al., 2006; DEMAT 5+: Götz et al.,
2013). The DEMAT has been used in several recent studies and showed high reliability and validity (e.g., Paetsch, Felbrich, & Stanat,
2015; Schuepbach, 2015; Winkelmann, Robitzsch, Stanat, & Köller, 2012). Each DEMAT version is designed for different grades and
includes three sections with (1) arithmetic, (2) geometry, and (3) word problems. By contrast to the IDS-2, children did not finish the
test based on performance but completed as many items as possible within a given time limit in each section. Correct answers were
summed up and this score was standardized with respect to each student’s grade level and used as a variable for our analyses. Re­
liabilities of these measures were high for each version of the DEMAT (Cronbach’s αs =.82-.91).

2.5. Executive functions (IDS-2)

The three components of EF, inhibition, cognitive flexibility, and working memory were measured separately using three tasks of

4
T. Kahl et al. Cognitive Development 62 (2022) 101149

the IDS-2. Inhibitory skills were assessed by using an adapted Stroop task (Stroop, 1935). Participants were asked to name the correct
color of four different animals (dolphin, chick, frog, and ladybug). This Stroop task and similar ones with, for instance, fruits have been
used in other recent studies (cf. Kahl et al., 2021; Neuenschwander, Röthlisberger, Cimeli, & Roebers, 2012; Röthlisberger, Neu­
enschwander, Cimeli, & Roebers, 2013). In three different conditions, participants were asked to name the correct color of each animal
presented as quickly and accurately as possible (for a total of 36 animals). The first condition was a set of congruently colored animals;
the second condition consisted of animals presented in black and white; the third condition consisted of incongruently colored animals.
The times of completion (t1, t2, and t3) and accuracy for each condition were measured. Based on previous studies (Archibald & Kerns,
1999; Röthlisberger, Neuenschwander, Michel, & Roebers, 2010), an inhibition score was created by computing these times according
to the formula (t3 – [t1xt2]/[t1+t2]).
To measure cognitive flexibility, participants were asked to name as many examples as possible that fit a given category. Previous
studies have used such verbal fluency tasks to assess cognitive flexibility in typically developing children (Stad et al., 2019; Van der
Elst, Hurks, Wassenberg, Meijs, & Jolles, 2011), children with developmental disorders (e.g., Leonard, Bernardi, Hill, & Henry, 2015)
and patients with frontal lobe lesions (e.g., Baldo, Shimamura, Delis, Kramer, & Kaplan, 2001). In accordance to Diamond (2013), this
task can be seen as a representative task assessing cognitive flexibility. Children aged 5 to 9 years were asked to name as many ex­
amples as possible within the categories “animals” and “food.” Children aged 10 to 20 years were additionally presented with the
categories “words beginning with E” and “words beginning with alternatingly S or L.” Time was limited to 90 s for each trial. The
number of correct words within target categories served as a cognitive flexibility score.
To measure working memory, participants were asked to recall different numbers or letters in forward and backward order as is
customary in other intelligence tests (e.g., WISC-IV; Wechsler, 2003). The task consisted of 40 items and started with an item con­
taining two digits or letters. To increase difficulty, additional digits or letters were added stepwise in each trial up to a maximum of
nine items to be recalled. The task was stopped after three subsequent incorrect recalls. The number of correctly recalled trials served
as a working memory score. Previous studies have used this kind of task to measure working memory (e.g., Chung, Liu, McBride, Wong,
& Lo, 2017; Fung et al., 2020).
These three components rely on each other and share variance (e.g., inhibition and cognitive flexibility require working memory
and vice versa; see Diamond, 2013). Therefore, we always accounted for the respective other components in our analyses. Scores were
age-standardized to reduce age-specific variance and to facilitate the comparability of scores in different age groups. Subtests
measuring EF showed satisfying reliabilities (Cronbach’s αs:.70-.75; Schmitt, 1996).

2.6. Visual-spatial skills (IDS-2)

Visual-spatial skills were measured using two different tasks that assessed children’s spatial visualization and spatial perception
(Linn & Peterson, 1985; Newcombe & Shipley, 2015). Both tasks have been used to operationalize visual-spatial skills in other studies
(e.g., Jirout & Newcombe, 2015; Kahl et al., 2021; Kyttälä et al., 2003; Markey, 2010). The task to measure spatial visualization
consisted of 20 items with increasing difficulty. Each item had a time limit (from 30-120 s). Children were asked to replicate
two-dimensional figures using triangular or rectangular cutouts (similar to the block design subtest in the WISC-IV; see Fig. 1). The task
was stopped after three subsequent false replications. The number of correctly replicated configurations was age-standardized and
used as a variable for our analyses.
In the task measuring spatial perception, children were presented with a printed pattern of circles and asked to copy this pattern
using thin plates on a separate sheet. The task consisted of four trials with increasingly difficult patterns. Each pattern had a time limit
of 60 seconds. Participant’s performance was operationalized as the accuracy of produced patterns as compared to the template. The
score ranged between zero (i.e., no overlap with the template) and two points (i.e., exact overlap). The number of points was summed
up. Again, age-standardized values were used for our analyses. Both visual-spatial tasks showed satisfying to high reliabilities for the
age groups (Cronbach’s αs:.72-.84)

2.7. Verbal reasoning

Two tasks were used to measure verbal reasoning; each subtest consisted of 34 items with increasing difficulty. In one subtest,
participants named the opposite of a given concept (e.g., rich vs. poor). The other subtest assessed how accurate participants named the

Fig. 1. In the spatial visualization task, participants were presented with 2-dimensional geometrical figures and asked to copy them using triangular
or rectangular cutouts.

5
T. Kahl et al. Cognitive Development 62 (2022) 101149

category of a set of different pictures (e.g., train, plane, bicycle: means of transportation). Each task was stopped when participants
gave three consecutive false answers. Neither task used a time limit. In both tasks, the number of correct answers were summed-up, and
a single, age-standardized composite score served as a variable for our analyses. Reliabilities for different age groups were high
(Cronbach’s αs:.86-.89).

Fig. 2. Differentiated cross-lagged panel model of executive functions, visual-spatial skills, and mathematical achievement. Participants (N = 182)
were 6 to 7 years old (6;00 - 7;11) at the time of their first assessment (Time point 1) and were re-assessed (n = 78) on average 3 years later at the
second time point (Time point 2). All manifest variables were age-standardized for analyses. We accounted for the variables sex, maternal education,
and verbal reasoning at T1. All variables correlated with each other. For the sake of clarity, the figure is divided into autoregressive effects (A) and
cross-lagged effects (B); however, both types of effects were computed statistically in one unified model. Furthermore, only significant paths (*p
<.05, **p <.001) and main variables have been included in the figure. Tdiff represents the time difference between T1 and T2.

6
T. Kahl et al. Cognitive Development 62 (2022) 101149

2.8. Socioeconomic status

We used the maternal educational level as a socioeconomic status index (cf. Hoff, 2013), as indicated on a five-point Likert scale
ranging from mandatory school education to university degree.

2.9. Data analytic procedure

We estimated a cross-lagged panel model (CLPM) using Mplus 7.0 (Muthén & Muthén, 2012) to assess the bidirectional, longi­
tudinal relationships among EF, visual-spatial skills, and mathematical performance (see Fig. 2). In such models, autoregressive paths
control for within-construct variance. Using this approach, it is possible to test multiple paths and examine the unique longitudinal
effects of several variables. Cross-lagged effects between constructs from prior to subsequent assessment time points indicate pre­
dictive effects of changes across time. We used a full information maximum likelihood (FIML) approach to handle missing values so
that all information could be used for our analysis. FIML is considered a robust method to deal with missing data in structural equation
modeling (SEM) and therefore also with CLPM (Cham, Reshetnyak, Rosenfeld, & Breitbart, 2017; Enders, 2001).
Since EF and visual-spatial skills were operationalized through various subtests, we aimed to model latent variables. However, the
bivariate correlations of the respective indicators of EF and visual-spatial skills were only moderately related to each other, and one
pair was not related at all (see Table 2). Similarly, when we computed principle-axis regression factor scores (Di Stefano, Zhu, &
Mîndrilă, 2009) to estimate the latent abilities of participants, we found they accounted for unsatisfactory portions of the variance (<
50%) in EF and visual-spatial skills. Hence, it was not possible to reliably model latent variables and an aggregated view of EF and
visual-spatial skills was not supported by the data. Consequently, we used a manifest approach and defined the CLPM using the
manifest variables of each subtest for a differentiated model (see Fig. 2). We accounted for the variables sex, maternal education, and
verbal reasoning at T1. Considering that the time differed between individual assessments among participants, we used
age-standardized values for all variables in our analyses that were measured at T1 and T2. To examine whether effects depended on
this time difference (Tdiff), we entered a time-difference variable into our analysis in addition to the interaction terms with this
respective variable and all main variables in a second step (e.g., working memory*Tdiff).
In SEM, power depends on a variety of specifics of the data and the model (Jak, Jorgensen, Verdam, Oort, & Elffers, 2020). Based on
a multitude of prior assumptions, simulations are conducted to determine necessary sample sizes. However, the model at hand can also
be considered as a combination of three regression models. The power calculation here requires fewer prior assumptions. We used
G*Power 3.1 to calculate a minimum sample size for a conventional longitudinal fixed effects regression analysis (Faul, Erdfelder,
Buchner, & Lang, 2009) with a medium incremental effect (f2 =.15), a statistical power of.80, and a two-tailed alpha probability level
of.01 (Cohen, 1988). Our calculation was based on an autoregressive model with six predictors, three control variables, one autore­
gressive variable and an interaction term. For such a model, a sample size of n = 83 would be required for sufficient power.

3. Results

Raw scores of the variables of interest at both assessment time points can be seen in Table 1. As expected, participants showed
higher average performance in all main variables at the second time point, which indicates developmental progression in each ability
tested. The following analyses of the current study were conducted using age-standardized values of the variables. Table 2 depicts
bivariate correlations including those among variables within each time point and between the assessment time points (cross-lagged
correlations). We found that mathematical achievement was significantly related to the variables working memory and spatial
visualization with rs ranging from.28 -.53. The variables inhibition, cognitive flexibility, and spatial perception showed significant but
rather small correlations with mathematical achievement at T1 (all rs <.26) and no significant correlations with mathematical
achievement at T2.
The differentiated CLPM can be seen in Fig. 2. The model is saturated given that there were as many estimated parameters as

Table 1
Descriptive statistics of the subtests at each measurement time point.
Time point 1 (N = 182) Time point 2 (n = 78) —————

Subtest M SD Min Max M SD Min Max Max score

Executive functions
Working memory 13.76 3.10 5 24 18.87 2.98 14 33 40
Inhibitiona 53.86 17.76 9 120 34.23 8.43 22 61 –
Cognitive flexibility 31.47 9.56 12 57 57.37 16.40 13 92 –
Visual-spatial skills
Spatial visualization 24.38 7.70 6 49 38.32 7.76 20 60 80
Spatial perception 12.74 4.39 1 23 26.60 6.75 9 41 76
Math. achievement (IDS-2) 18.15 5.76 7 36 33.88 6.65 22 50 64
Math. achievement (DEMAT) – – – – 54.44 9.02 34 80 –

Note. Children were 6 to 7 years old (6;00 - 7;11) at the time of their first assessment (Time point 1) and were re-assessed on average 3 years later
(Time point 2).
a
Values of the inhibition task are reversed so that lower values indicate higher inhibitory skills.

7
 T. Kahl et al.
Table 2
Bivariate correlations among variables of interest at both assessment time points (Time point 1: N = 182; Time point 2: n = 78).
Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Time point 1
1 Sex -
2 Maternal education .083 -
3 Verbal reasoning .082 .293* -
4 Working memory -.023 .175* .446** -
5 Inhibition .196 .166 .201* .303** -
6 Cognitive flexibility .148 .159 .410** .276** .222** -
7 Spatial visualization -.158* .278 .468** .401* .265* .231** -
8 Spatial perception -.047 .028 .229** .261 .143 .120 .376** -
8

9 Math. achievement (IDS-2) -.235** .285** .512** .534** .239** .264** .442** .250** -
Time point 2
10 Working memory -.113 .262* .291** .555** .103 -.012 .191 .093 .463** -
11 Inhibition -.038 .094 .015 .031 .132 .077 .131 -.079 .113 .171 -
12 Cognitive flexibility .207 .208 .188 .078 .069 .414** .156 .240* .109 .235* .253* -
13 Spatial visualization -.240* .275* .199 .329** .196 .078 .605 .346** .394** .330** .187 .161 -
14 Spatial perception -.012 -.017 -.098 .071 -.025 .039 .180 .024 .020 .037 .177 .085 .283* -
15 Math. achievement (IDS-2) -.380** .183 .336** .293** .041 .136 .334** .200 .575** .385** .026 .205 .474** .072 -
16 Math. achievement (DEMAT) -.262* .375** .142 .282* -.002 .021 .331** .161 .452** .300** .248* .247* .496** .181 .557** -

Note. IDS-2 = Intelligence and Development Scales-2; DEMAT = Deutscher Mathematiktest (The German Test of Mathematical Abilities). Age-standardized values were used for these correlations.
*p <.05, **p <.00

Cognitive Development 62 (2022) 101149

T. Kahl et al. Cognitive Development 62 (2022) 101149

observed statistics (variances and covariances) in the data set. All variables correlate to each other in this model. We accounted for sex,
maternal education, and verbal reasoning by regressing predictor variables at T1 and criteria variables at T2 on these variables.
Autoregressive effects of variables between the two assessment time points indicate the degree of stability of the constructs over time
(Selig & Little, 2012). Our CLPM indicates stability across the 3-year period for the following variables: mathematical achievement
measured with the IDS-2: β =.496, p <.001, mathematical achievement measured with the DEMAT: β =.396, p <.001, working
memory: β =.480, p <.001, cognitive flexibility: β =.359, p <.001, and spatial visualization: β =.513, p <.001. Inhibition (β =.147, p
=.22) and spatial perception (β = -.041, p =.75) do not indicate stability between the two assessments even though spatial perception
and spatial visualization were strongly related across these time points.
With respect to cross-sectional relations, the analyses showed that mathematical achievement (DEMAT) was predicted by spatial
visualization, β =.247, p =.03, and by children’s earlier mathematical achievement (IDS-2), β =.396, p =.002. Therefore, it seems that
early spatial visualization predicts later mathematical performance, even after accounting for earlier mathematical ability, EF, sex,
maternal education, and verbal reasoning. However, given that the mathematical achievement score of the DEMAT contains some
geometry items, we conducted another analysis in which only the “word problems” section of the DEMAT served as an indicator of
mathematical performance. The results show an identical predictive effect by spatial visualization with an even stronger beta coef­
ficient, β =.275, p =.02. No other indicator of EF or the other visual-spatial skill (spatial perception) directly predict children’s later
mathematical achievement, although a trend is observed for working memory (β =.175, p =.09).
With respect to investigating the effects of the time difference between T1 and T2 (Tdiff), results show that the time difference did
not predict later mathematical achievement (β = -.041, p =.67). Therefore, time intervals between assessments are not associated with
participants’ mathematical achievement. However, we found that the time difference moderates the relation between working
memory at T1 and changes in later mathematical achievement. This is indicated by a significant interaction term Working Memo­
ry*Tdiff (β =.285, p <.001), whereas no other interaction term reaches significance. Looking at the simple slope analyses (see Fig. 3),
we found that the further apart the assessments were, the more children’s mathematical achievement changed in those children who
scored high on working memory at T1 (slope: β =.37, p <.001). When assessments were closer to each other, Fig. 3 implies that
children’s mathematical achievement changed more in children scoring low on working memory at T1; however, this effect is only a
tendency (slope: β = -.23, p =.064).
With respect to the reversed directions, we found that children’s earlier mathematical abilities as measured with the IDS-2 predict
changes in working memory at T2, β =.306, p =.017. Therefore, it seems that children’s early mathematical ability is predictive of their
later working memory, even after accounting for their earlier working memory, sex, maternal education, and verbal reasoning. By
contrast, children’s earlier mathematical abilities are un-related to their visual-spatial skills at T2. Finally, performance on spatial
perception at T1 predict changes in cognitive flexibility at T2, β =.201, p =.026.

4. Discussion

In the current study, we investigated bidirectional associations among EF (working memory, inhibition, and cognitive flexibility),
visual-spatial skills, and mathematical achievement in primary school children. We used a longitudinal design with a sample of 182
children, tested at the age of 6 to 7 years and again 3 years later. Our results indicate that early spatial visualization predicts math­
ematical achievement 3 years later and that early working memory predicts changes in mathematical achievement, depending on the
time difference between the assessment time points. Furthermore, early mathematical achievement predicts later changes in working
memory.
Previous research has shown that early EF and visual-spatial skills predict mathematical achievement at a later time point (e.g., EF:
Clark et al., 2010; visual-spatial skills: Verdine et al., 2014). Investigating such unidirectional associations reflects the predominant

Fig. 3. Changes in participants’ mathematical achievement (standardized residuals) as a function of individual differences in working memory at T1
(z-Scores). Separate lines are shown for participants being tested one standard deviation above (Tdiff – large) or below (Tdiff – short) average time
differences (Tdiff) between T1 and T2.

9
T. Kahl et al. Cognitive Development 62 (2022) 101149

line of research so far. Only a few studies have investigated bidirectional relations between cognitive skills and academic performance;
however, most of these studies investigated EF and visual-spatial skills separately and mainly in preschool or kindergarten children.
The results of these studies indicate bidirectional patterns between EF, visual-spatial skills, and math in children prior to primary
school (e.g., Fuhs et al., 2014; Geer et al., 2019, but see Fung et al., 2020). Our study has extended this research by using a cross-lagged
design to investigate bidirectional relations among EF, visual-spatial skills, and mathematical achievement in primary school students.
Research on this specific group seems especially relevant, as primary school students are introduced to formal mathematics and most
psychological assessments and interventions take place during this time.
We have found that spatial visualization predicts mathematical achievement 3 years later during primary school. Children who
scored higher on spatial visualization in first or second grade also scored higher in a standardized math test 3 years later, even after
accounting for their mathematical achievement at T1, EF, maternal education, and verbal reasoning. However, our data have not
yielded evidence for the reversed relation with earlier mathematical achievement predicting children’s later spatial visualization,
which indicates a unidirectional relationship between these variables. This asymmetrical result is in line with findings from Fung et al.
(2020), but contrasts with findings of another recent study showing bidirectional relations (Geer et al., 2019). A potential explanation
may be that the current study and the one from Fung and colleagues controlled for a number of additional predictors of mathematical
achievement such as EF, maternal education, or verbal intelligence, whereas the study from Geer et al. (2019) did not. Therefore, it is
possible that other co-occurring variables included in the performance of a visual-spatial task (such as EF) may have contributed to the
bidirectional result from Geer et al. (2019). In other words, the impurity problem in measuring constructs may result in apparent
bidirectional relations between constructs that we have accounted for in our model.
Overall, our research highlights that spatial visualization seems to play a specific role in developing mathematical skills. However,
the underlying mechanisms of the relation between visual-spatial skills and mathematics are still hotly debated. Some researchers
think that visual-spatial skills might help with visualizing mathematical problems, which is beneficial to solving them (Hawes &
Ansari, 2020; Hegarty & Kozhevnikov, 1999). Other researchers advocate that spatial skills may help with imagining numerical
magnitude and ordering this magnitude on a mental number line (cf. Gunderson et al., 2012), which is a foundational concept for
learning mathematics (Ramani & Siegler, 2008; Schneider et al., 2018). In addition to these ideas about mechanisms, relations may
also depend on the kinds of visual-spatial skill and mathematical topics being measured. Our data implies no association between
mathematical achievement and spatial perception––the second measure of visual-spatial skills in our study. This finding is in line with
recent results (Geer et al., 2019), though we did find associations for mathematical achievement and spatial visualization. A recent
meta-analysis by Xie and colleagues (2020) has shown that the strength of the association between visual-spatial skills and mathe­
matics depends on the mathematical areas investigated. For instance, logical reasoning showed a stronger relation to visual-spatial
skills than to arithmetic. In our study, we show similar effects for two different areas, namely a curriculum-based measure
comprising various school-related topics and the more specific measure of solving word problems.
With respect to relations between EF and mathematical achievement, we have found that working memory predicts changes in
mathematical achievement, depending on the time difference between the two assessments. Our findings indicate that children with
high working memory skills were at an advantage in mathematical achievement when assessments were further apart, whereas a trend
indicates that children with low working memory skills improved more in mathematical achievement when assessments were closer to
each other. This finding may indicate that the initial advantages of children with high working memory at T1 prevail and increase in
the long term, whereas they might be relativized in the short term by catch-up effects from children with low working memory skills.
One reason may be children’s differential use of strategies. Children with high working-memory abilities may use more advanced
strategies that also require more working memory (DeStefano & LeFevre, 2004; Geary, Hoard, Byrd-Craven, & DeSoto, 2004; Ramirez,
Gunderson, Levin, & Beilock, 2012). Using more elaborate strategies (e.g., decomposition, retrieval) demands high working memory
and is related to higher conceptual understanding and mathematical achievement (Barrouillet & Lépine, 2005; Mazzocco, Devlin, &
McKenney, 2008). This may explain why children with high working memory skills improve in their mathematical achievement over a
longer time period as more elaborate strategies need more time to pay off. After longer time periods students may also be confronted
with more complex mathematical content and may benefit more from their higher-order strategies. In contrast, children with lower
capacities may use easier heuristics and invest more effort in general (Geary et al., 2004; Fung et al., 2020; Ramirez, Chang, Maloney,
Levine, & Beilock, 2016). Notably, this effect is specific for working memory and does not appear for any other component of EF. It
highlights the important role of working memory for mathematical learning (e.g., Cragg & Gilmore, 2014; Geary, Hoard, Byrd-Craven,
Nugent, & Numtee, 2007).
Importantly, our results also indicate reversed effects, suggesting that earlier mathematical achievement predicts changes in
working memory 3 years later. This result is in line with other studies of children in kindergarten or preschool (e.g., Fuhs et al., 2014;
Miller-Cotto & Byrnes, 2019; Schmitt et al., 2017). A possible mechanism for this reversed effect might be the reciprocal and beneficial
relation between learning mathematics and working memory capacities. As primary school children are frequently exposed to
mathematics, they have longer learning times and this exposure to mathematics may train children’s working memory simultaneously
(Fuhs et al., 2014). This assumption is strengthened by the fact that working memory trainings often consist of mathematical tasks (for
an overview, Diamond, 2012). A theoretical explanation for these bidirectional effects between working memory and mathematical
achievement can be found in theories of the transfer of learning (Klauer, 2000) and the transactional model (Peng & Kievit, 2020).
According to these theories, domain-general skills (e.g., working memory) are automatically trained when specific skills are trained (e.
g., mathematical skills) and vice versa.
The current study has both strengths and limitations. It is a strength that we investigate bidirectional associations between
mathematical achievement and two important domain-general skills, EF and visual-spatial skills, in a single model. Furthermore, we
are the first to investigate these relations in primary school students––a group extensively exposed to formal mathematics. Moreover,

10
T. Kahl et al. Cognitive Development 62 (2022) 101149

all our participants were assessed using standardized tests. Mathematical achievement was measured using two tests, both based on
children’s school curriculum but with slightly different topics and procedures. At the same time, our study shows some limitations that
should be considered when interpreting the findings. Given that no reliable latent variables could be modeled, it was impossible to
provide a parsimonious model of EF and visual spatial skills. Relatedly, assessment effects cannot be excluded. The pattern of results
may have been influenced by the psychometric properties of the test procedures used. At the same time, this approach is in line with the
current state of knowledge according to which both executive functions and visual spatial skills are umbrella terms of moderately
correlated and yet separable skills. Latent modeling or aggregation into an index score would have led at best to a loss of information
and at worst to a mis-specified model. However, ideally, future research may assess each variable through various measures (e.g., three
different measures for cognitive flexibility) in order to use a reliable latent variable approach.
Furthermore, even though we used a bidirectional approach, a causal perspective on these relations was not possible. The co­
efficients must be interpreted with caution given that a CLPM with two assessment time points does not allow distinguishing between
the within- and between-person levels (Bailey, Oh, Farkas, Morgan, & Hillemeier, 2019; Hamaker, Kuiper, & Grasman, 2015) and the
mutual effects might be inflated (Curran & Hancock, 2021; see also Feldon & Litson, 2021). Finally, the drop-out analyses showed that
the remaining subsample was rather selective with higher scores on all main variables and verbal reasoning. This selection bias is
linked to the recruitment of participants at T2. Even though the entire subsample (N = 182) was invited, children showing higher
cognitive and mathematical abilities were more responsive to participating again at T2. Therefore, generalizability of the findings is
limited and should be considered when interpreting the present results.
To date, research on influencing variables of mathematical achievement has focused especially on unidirectional assumptions.
However, the last decade of research has shown that patterns between cognitive and academic domains seem to interact and develop
mutually. Following this new line of research, the present longitudinal study investigated bidirectional associations between domain-
general skills and mathematical achievement. Our results indicate at least some cognitive and mathematical skills develop in tandem
and interact mutually across child development. Our results entail important information for intervention and training studies for
primary school students as they show that academic and cognitive development can be reciprocal and co-develop beneficially.
Therefore, reciprocal relations may be used in interventions by for instance simultaneously training working memory and mathe­
matical skills. Such combined trainings revealed rather promising results as mentioned in a recent review (Hawes & Ansari, 2020). A
careful manipulation in which such combined trainings are systematically compared to single trainings will inform about underlying
causal relations which may shed light on these dynamics and allow for specific suggestions for mathematics curricula, interventions
and assessments.

Declaration of Conflicting Interests

The authors declare that they had no conflicts of interest with respect to their authorship or the publication of this article.

Acknowledgements

A special thank goes to the Suzanne and Hans Biaesch Foundation for the Enhancement of Applied Psychology (Switzerland) for
funding the current research project. Further, we are grateful to our colleagues of the Division of Developmental and Personality
Psychology for their helpful feedback during the brown bag meetings. In particular, we wish to thank the research assistants who were
in charge of data collection.

References

Allan, N. P., Hume, L. E., Allan, D. M., Farrington, A. L., & Lonigan, C. J. (2014). Relations between inhibitory control and the development of academic skills in
preschool and kindergarten: A meta-analysis. Developmental Psychology, 50(10), 2368–2379. https://doi:10.1037/a0037493.
Archibald, S. J., & Kerns, K. A. (1999). Identification and description of new tests of executive functioning in children. Child Neuropsychology, 5, 115–129.
Baddeley, A. (1996). Exploring the central executive. The Quarterly Journal of Experimental Psychology Section, 49A(1), 5–28. https://doi.org/10.1080/713755608.
Bailey, D. H., Oh, Y., Farkas, G., Morgan, P., & Hillemeier, M. (2019). Reciprocal effects of reading and mathematics? beyond the cross-lagged panel model.
Developmental Psychology, 56(5), 912–921. https://doi.org/10.1037/dev0000902
Baldo, J. V., Shimamura, A. P., Delis, D. C., Kramer, J., & Kaplan, E. (2001). Verbal and design fluency in patients with frontal lobe lesions. Journal of the International
Neuropsychological Society, 7, 586–596.
Blair, C., & Razza, R. P. (2007). Relating effortful control, executive function, and false belief understanding to emerging math and literacy ability in kindergarten.
Child Development, 78(2), 647–663. https://doi.org/10.1111/j.1467-8624.2007.01019.x
Barrouillet, P., & Lépine, R. (2005). Working memory and children’s use of retrieval to solve addition problems. Journal of Experimental Child Psychology, 91(3),
183–204. https://doi.org/10.1016/j.jecp.2005.03.002
Bull, R., & Scerif, G. (2001). Executive functioning as a predictor of children’s mathematics ability: Inhibition, switching, and working memory. Developmental
Neuropsychology, 19(3), 273–293. https://doi.org/10.1207/S15326942DN1903_3
Butterworth, B. (2005). The development of arithmetical abilities. Journal of Child Psychology and Psychiatry, 46(1), 3–18. https://doi.org/10.1111/j.1469-
7610.2004.00374.x
Casey, M., Nuttall, R., & Benbow, C. (1995). The influence of spatial ability on gender differences in mathematics college entrance test-scores across diverse samples.
Developmental Psychology, 31(4), 697–705. https://doi.org/10.1037/0012-1649.31.4.697
Cattell, R. B. (1987). Intelligence: Its structure, growth and action. Elsevier.
Cham, H., Reshetnyak, E., Rosenfeld, B., & Breitbart, W. (2017). Full information maximum likelihood estimation for latent variable interactions with incomplete
indicators. Multivariate Behavioral Research, 52, 12–30. https://doi.org/10.1080/ 00273171.2016.1245600.
Chu, F. W., van Marle, K., & Geary, D. (2016). Predicting children’s reading and mathematicsachievement from early quantitative knowledge and domain-general
cognitive.abilities. Frontiers in Psychology, 7, 775. https://doi.org/10.3389/fpsyg.2016.00775

11
T. Kahl et al. Cognitive Development 62 (2022) 101149

Chung, K. K., Liu, H., McBride, C., Wong, A. M. Y., & Lo, J. C. (2017). How socioeconomic status, executive functioning and verbal interac- tions contribute to early
academic achievement in Chinese children. Educational Psychology, 37, 402–420. https://doi.org/10.1080/ 01443410.2016.1179264.
Clark, C. A., Pritchard, V. E., & Woodward, L. J. (2010). Preschool executive functioning abilities predict early mathematics achievement. Developmental Psychology, 46
(5), 1176–1191. https://doi.org/10.1037/a0019672
Cohen. (1988). Statistical power analysis for the behavioral sciences ((2nd ed.)). Lawrence Erlbaum Associates.
Cragg, L., & Gilmore, C. (2014). Skills underlying mathematics: The role of executive function in the development of mathematics proficiency. Trends in Neuroscience
and Education, 3(2), 63–68. https://doi.org/10.1016/j.tine.2013.12.001
Cragg, L., Keeble, S., Richardson, S., Roome, H. E., & Gilmore, C. (2017). Direct and indirect influences of executive functions on mathematics achievement. Cognition,
162, 12–26. https://doi.org/10.1016/j.cognition.2017.01.014
Curran, P. J., & Hancock, G. R. (2021). The challenge of modeling co-developmental processes over time. Child Development Perspectives, 15(2), 67–75.
Dehaene, S. (1992). Varieties of numerical abilities. Cognition, 44(1–2), 1–42. https://doi.org/10.1016/0010-0277(92)90049-N
DeStefano, D., & LeFevre, J. A. (2004). The role of working memory in mental arithmetic. European Journal of Cognitive Psychology, 16, 353–386. https://doi.org/
10.1080/09541440244000328
Diamond, A. (2012). Activities and programs that improve children’s executive functions. Current Directions in Psychological Science, 21(5), 335–341. https://doi.org/
10.1177/0963721412453722
Diamond, A. (2013). Executive functions. Annual Review of Psychology, 64, 135–168. https://doi.org/10.1146/annurev-psych-113011-143750
Di Stefano, C., Zhu, M., & Mîndrilă, D. (2009). Understanding and using factor scores: Considerations for the applied researcher. Practical Assessment. Research &
Evaluation, 14(20), 1–11.
Doebel, S. (2020). Rethinking executive function and its development. Perspectives on Psychological Science, 15(4), 942–956. https://doi.org/10.1177/
1745691620904771
Duncan, G. J., Dowsett, C. J., Claessens, A., Magnuson, K., Huston, A. C., Klebanov, P., … Japel, C. (2007). School readiness and later achievement. Developmental
Psychology, 43(6), 1428–1446. https://doi.org/10.1037/0012-1649.43.6.1428
Dunning, D. L., Holmes, J., & Gathercole, S. E. (2013). Does working memory training lead to generalized improvements in children with low working memory? A
randomized controlled trial. Developmental Science, 16(6), 915–925. https://doi.org/10.1111/desc.12068
Enders, C. K. (2001). The impact of nonnormality on full information maximum-likelihood estimation for structural equation models with missing data. Psychological
Methods, 6, 352–370. https://doi.org/10.1037/1082-989X.6.4.352
Erziehungsdirektoren-Konferenz (D-EDK). (2013). Lehrplan 21. Mathematik (Konsultationsfassung). Luzern: D-EDK.
Feldon, D. F., & Litson, K. (2021). Modeling theories and theorizing models: an attempted replication of Miller-Cotto & Byrnes’(2019) comparison of working memory
models using ECLS-K data. Educational Psychology Review, 1–28.
Frick, A. (2019). Spatial transformation abilities and their relation to later mathematics performance. Psychological Research, 83, 1465–1484. https://doi.org/
10.1007/s00426-018-1008-5
Friso-Van Den Bos, I., der Ven, Van, Kroesbergen, S. H., H., E., Luit, Van, & J. E.. (2013). Working memory and mathematics in primary school children: a meta-
analysis. Educational Research Review, 10, 29–44. https://doi.org/10.1016/j.edurev.2013.05.003
Fuhs, M. W., Nesbitt, K. T., Farran, D. C., & Dong, N. (2014). Longitudinal associations between executive functioning and academic skills across content areas.
Developmental Psychology, 50(6), 1698–1709. https://doi.org/10.1037/a0036633
Fung, W. K., Chung, K. K. H., & Lam, C. B. (2020). Mathematics, executive functioning, and visual-spatial skills in Chines kindergarten children: Examining the
bidirectionality. Journal of Experimental Child Psychology, 199, 1–10. https://doi.org/10.1016/j.jecp.2020.104923
Geary, D. C. (2000). From infancy to adulthood: the development of numerical abilities. European Child & Adolescent Psychiatry, 9, 11–16. https://doi.org/10.1007/
s007870070004
Geary, D. C. (2011). Cognitive predictors of achievement growth in mathematics: a five year longitudinal study. Developmental Psychology, 47(6), 1539–1552. https://
doi.org/10.1037/a0025510
Geary, D. C., Hoard, M. K., Byrd-Craven, J., & DeSoto, M. C. (2004). Strategy choice in simple and complex addition: Contributions of working memory and counting
knowledge for children with mathematical disability. Journal of Experimental Child Psychology, 88, 121–151. https://doi.org/10.1016/j.jecp.2004.03.002
Geary, D. C., Hoard, M. K., Byrd-Craven, J., Nugent, L., & Numtee, C. (2007). Cognitive mechanisms underlying achievement deficits in children with mathematical
learning disability. Child Development, 78(4), 1343–1359. https://doi.org/10.1111/j.1467-8624.2007.01069.x
Geer, E. A., Quinn, J. M., & Ganley, C. M. (2019). Relations between spatial skills and math performance in elementary school children: A longitudinal investigation.
Developmental Psychology, 55(3), 637–652. https://doi.org/10.1037/dev0000649
Gölitz, D., Roick, T., & Hasselhorn, M. (2006). Deutscher Mathematiktest für vierte Klassen (DEMAT 4). Göttingen: Hogrefe.
Götz, L., Lingel, K., & Schneider, W. (2013). Deutscher Mathematiktest für fünfte Klassen (DEMAT 5). Göttingen: Hogrefe.
Grob, A. & Hagmann-von Arx, P. (2018). Intelligence and developmental scales – 2 (IDS-2). Hogrefe.
Gunderson, E. A., Ramirez, G., Beilock, S. L., & Levine, S. C. (2012). The relation between spatial skill and early number knowledge: The role of the linear number line.
Developmental Psychology, 48(5), 1229–1241. https://doi.org/10.1037/a0027433
Hamaker, E. L., Kuiper, R. M., & Grasman, R. P. (2015). A critique of the cross-lagged panel model. Psychological Methods, 20(1), 102. https://doi.org/10.1037/
a0038889
Hawes, Z., & Ansari, D. (2020). What explains the relationship between spatial and mathematical skills? A review of evidence from brain and behavior. Psychonomic
Bulletin & Review, 1–18. https://doi:10.3758/s13423-019-01694-7.
Hegarty, M., & Kozhevnikov, M. (1999). Types of visual-spatial representations and mathematical problem solving. Journal of Educational Psychology, 91, 684–689.
https://doi.org/10.1037/0022-0663.91.4.684
Hoff, E. (2013). Interpreting the early language trajectories of children from low-SES and language minority homes: implications for closing achievement gaps.
Developmental Psychology, 49(1), 4. https://doi.org/10.1037/a0027238
Jak, S., Jorgensen, T. D., Verdam, M. G., Oort, F. J., & Elffers, L. (2020). Analytical power calculations for structural equation modeling: A tutorial and Shiny app.
Behavior Research Methods, 1–22.
Jirout, J. J., & Newcombe, N. S. (2015). Building blocks for developing spatial skills: Evidence from a large, representative US sample. Psychological Science, 26(3),
302–310. https://doi.org/10.1177/0956797614563338
Kahl, T., Grob, A., Segerer, R., & Möhring, W. (2021). Executive functions and visual-spatial skills predict mathematical achievement: asymmetrical associations
across age. Psychological Research, 85, 36–46. https://doi.org/10.1007/s00426-019-01249-4
Klauer, K. J. (2000). Das Huckepack-Theorem asymmetrischen Strategietransfers. Ein Bei- trag zur Trainings- und Transfertheorie. Zeitschrift für
Entwicklungspsychologie und Pädagogische Psychologie, 32(3), 153–165.
Krajewski, K., Liehm, S., & Schneider, W. (2004). Deutscher Mathematiktest für zweite Klassen (DEMAT 2+). Beltz.
Kubesch, S. (2016). Exekutive Funktionen und Selbstregulation: Neurowissenschaftliche Grundlagen und Transfer in die pädagogische Praxis. Hogrefe.
Kuncel, N. R., Hezlett, S. A., & Ones, D. S. (2004). Academic performance, career potential, creativity, and job performance: Can one construct predict them all?
Journal of Personality and Social Psychology, 86(1), 148–161. https://doi.org/10.1037/0022-3514.86.1.148
Kyttälä, M., Aunio, P., Lehto, J. E., Van Luit, J., & Hautamaki, J. (2003). Visuospatial working memory and early numeracy. Educational and Child Psychology, 20(3),
65–76.
Lee, K., & Bull, R. (2016). Developmental changes in working memory, updating, and math achievement. Journal of Educational Psychology, 108(6), 869–882. https://
doi.org/10.1037/edu0000090
Lee, K., Ng, S. F., Pe, M. L., Ang, S. Y., Hasshim, M. N. A. M., & Bull, R. (2012). The cognitive underpinnings of emerging mathematical skills: Executive functioning,
patterns, numeracy, and arithmetic. British Journal of Educational Psychology, 82(1), 82–99. https://doi.org/10.1111/j.2044-8279.2010.02016.x

12
T. Kahl et al. Cognitive Development 62 (2022) 101149

Leonard, H. C., Bernardi, M., Hill, E. L., & Henry, L. A. (2015). Executive functioning, motor difficulties, and developmental coordination disorder. Developmental
Neuropsychology, 40(4), 201–215.
Lindberg, S. M., Hyde, J. S., Petersen, J. L., & Linn, M. C. (2010). New trends in gender and mathematics performance: A meta-analysis. Psychological Bulletin, 136(6),
1123–1135. https://doi:10.1037/a0021276.
Linn, M. C., & Petersen, A. C. (1985). Emergence and characterization of sex differences in spatial ability: A meta-analysis. Child Development, 56(6), 1479–1498.
https://doi.org/10.1111/j.1467- 8624.1985.tb00213.x
Lyons, M. D., & Huebner, E. S. (2016). Academic characteristics of early adolescents with higher levels of life satisfaction. Applied Research Quality Life, 11, 757–771.
https://doi.org/10.1007/s11482-015-9394-y
Markey, S. M. (2010). The relationship between visual-spatial reasoning ability and math and geometry problem solving. Dissertation Abstracts International: Section B: The
Sciences and Engineering, 70, 7874.
Mazzocco, M. M., Devlin, K. T., & McKenney, S. J. (2008). Is it a fact? Timed arithmetic performance of children with mathematical learning disabilities (MLD) varies
as a function of how MLD is defined. Developmental Neuropsychology, 33(3), 318–344. https://doi.org/10.1080/87565640801982403
Miller-Cotto, D., & Byrnes, J. P. (2019). What’s the best way to characterize the relationship between working memory and achievement? An initial examination of
competing theories. Journal of Educational Psychology. https://doi.org/10.1037/edu0000395
Mix, K. S., & Cheng, Y. L. (2012). The relation between space and math: developmental and educational implications. In J. B. Benson (Ed.), Advances in child
development and behavior, 42 pp. 197–243). Elsevier Academic Press Inc.
Miyake, A., & Friedman, N. P. (2012). The nature and organization of individual differences in executive functions. Current Directions in Psychological Science, 21(1),
8–14. https://doi.org/10.1177/0963721411429458
Möhring, W., Frick, A., & Newcombe, N. S. (2018). Spatial scaling, proportional thinking, and numerical understanding in 5- to 7-year-old children. Cognitive
Development, 45, 57–67. https://doi.org/10.1016/j.cogdev.2017.12.001
Monette, S., Bigras, M., & Guay, M. C. (2011). The role of the executive functions in school achievement at the end of Grade 1. Journal of Experimental Child Psychology,
109(2), 158–173. https://doi.org/10.1016/j.jecp.2011.01.008
Muthén, L. K., & Muthén, B. O. (2012). Mplus user’s guide (7th ed.). Muthén & Muthén.
Neuenschwander, R., Röthlisberger, M., Cimeli, P., & Roebers, C. M. (2012). How do different aspects of self-regulation predict successful adaptation to school?
Journal of Experimental Child Psychology, 113(3), 353–371. https://doi.org/10.1016/j.jecp.2012.07.004
Newcombe, N. S., Möhring, W., & Frick, A. (2018). How big is many? Development of spatial and numerical magnitude understanding. In A. Henik, & W. Fias (Eds.),
Heterogeneity of function in numerical cognition (pp. 157–176). Elsevier Academic Press.
Newcombe, N. S., & Shipley, T. F. (2015). Thinking about spatial thinking: new typology, new assessments. Studying visual and spatial reasoning for design creativity (pp.
179–192). Springer. https://doi.org/10.1007/978-94-017-9297-4_10
Paetsch, J., Felbrich, A., & Stanat, P. (2015). Der Zusammenhang von sprachlichen und mathematischen Kompetenzen bei Kindern mit Deutsch als Zweitsprache.
Zeitschrift für Pädagogische Psychologie, 29(1), 19–29.
Paglin, M., & Rufolo, A. M. (1990). Heterogeneous human capital, occupational choice, and male–female earnings differences. Journal of Labor Economics, 8, 123–144.
https://doi.org/10.1086/298239
Peng, P., & Goodrich, J. M. (2020). The cognitive element model of reading instruction. Reading Research Quarterly, 55, S77–S88.
Peng, P., & Kievit, R. A. (2020). The development of academic achievement and cognitive abilities: A bidirectional perspective. Child Development Perspectives, 14(1),
15–20. https://doi.org/10.1111/cdep.12352
Peng, P., Wang, T., Wang, C., & Lin, X. (2019). A meta-analysis on the relation between fluid intelligence and reading/mathematics: Effects of tasks, age, and social
economics status. Psychological Bulletin, 145, 189–236. https://doi.org/10.1037/bul0000182
Ramani, G. B., & Siegler, R. S. (2008). Promoting broad and stable improvements in low-income children’s numerical knowledge through playing number board
games. Child Development, 79(2), 375–394. https://doi.org/10.1111/j.1467-8624.2007.01131.x
Ramirez, G., Chang, H., Maloney, E. A., Levine, S. C., & Beilock, S. L. (2016). On the relationship between math anxiety and math achievement in early elementary
school: The role of problem solving strategies. Journal of Experimental Child Psychology, 141, 83–100. https://doi.org/10.1016/j. jecp.2015.07.014
Ramirez, G., Gunderson, E. A., Levine, S. C., & Beilock, S. L. (2012). Spatial anxiety relates to spatial abilities as a function of working memory in children. Quarterly
Journal of Experimental Psychology, 65(3), 474–487. https://doi.org/10.1080/17470218.2011.616214
Rivera-Batiz, F. L. (1992). Quantitative literacy and the likelihood of employment among young adults in the United States. Journal of Human Resources, 27(2),
313–328. 〈http://www.jstor.org/stable/pdfplus/145737〉.
Roick, T., Gölitz, D., & Hasselhorn, M. (2004). Deutscher Mathematiktest für dritte Klassen (DEMAT 3). Hogrefe.
Röthlisberger, M., Neuenschwander, R., Cimeli, P., & Roebers, C. M. (2013). Executive functions in 5-to 8-year olds: Developmental changes and relationship to
academic achievement. Journal of Educational and Developmental Psychology, 3, 153–167. https://doi.org/10.5539/jedp.v3n2p153
Röthlisberger, M., Neuenschwander, R., Michel, E., & Roebers, M. (2010). Exekutive Funktionen: Zugrundeliegende kognitive Prozesse und deren Korrelate bei
Kindern im späten Vorschulalter. Zeitschrift für Entwicklungspsychologie und Pädagogische Psychologie, 42(2), 99–110.
Schmitt, N. (1996). Uses and abuses of coefficient alpha. Psychological Assessment, 8(4), 350–353. https://doi.org/10.1037/1040-3590.8.4.350
Schmitt, S. A., Geldhof, G. J., Purpura, D. J., Duncan, R., & McClelland, M. M. (2017). Examining the relations between executive function, math, and literacy during
the transition to kindergarten: A multi-analytic approach. Journal of Educational Psychology, 109(8), 1120. https://doi.org/10.1037/edu0000193
Schneider, M., Merz, S., Stricker, J., De Smedt, B., Torbeyns, J., Verschaffel, L., & Luwel, K. (2018). Associations of number line estimation with mathematical
competence: A meta-analysis. Child development, 89(5), 1467–1484. https://doi.org/10.1111/cdev.13068
Schuepbach, M. (2015). Effects of extracurricular activities and their quality on primary school-age students’ achievement in mathematics in Switzerland. School
Effectiveness and School Improvement, 26(2), 279–295. https://doi.org/10.1080/09243453.2014.929153
Selig, J. P., & Little, T. D. (2012). Autoregressive and cross-lagged panel analysis for Longitudinal data. In B. Laursen, & T. D. Little (Eds.), Handbook of developmental
research methods (pp. 265–278). Guilford Press.
Stad, F. E., Wiedl, K. H., Vogelaar, B., Bakker, M., & Resing, W. C. M. (2019). The role of cognitive flexibility in young children’s potential for learning under dynamic
testing conditions. European Journal of Psychology of Education, 34, 123–146, 10.1007/s10212-018-0379-8.
Sternberg, R. J., Kaufman, J. C., & Grigorenko, E. L. (2008). Applied intelligence. Cambridge University Press.
Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18, 643–662. https://doi:10.1037/h0054651.
Uttal, D. H., Meadow, N. G., Tipton, E., Hand, L. L., Alden, A. R., Warren, C., & Newcombe, N. S. (2013). The malleability of spatial skills: A meta-analysis of training
studies. Psychological Bulletin, 139, 352–402. https://doi.org/10.1037/a0028446
Van der Elst, W., Hurks, P., Wassenberg, R., Meijs, C., & Jolles, J. (2011). Animal verbal fluency and design fluency in school-aged children: Effects of age, sex, and
mean level of parental education, and regression-based normative data. Journal of Clinical and Experimental Neuropsychology, 33(9), 1005–1015. https://doi.org/
10.1080/13803395.2011.589509
Vandenbroucke, L., Verschueren, K., & Baeyens, D. (2017). The development of executive functioning across the transition to first grade and its predictive value for
academic achievement. Learning and Instruction, 49, 103–112. https://doi.org/10.1016/j.learninstruc.2016.12.008
van der Maas, H. L. J., Dolan, C. V., Grasman, R. P. P. P., Wicherts, J. M., Huizenga, H. M., & Raijmakers, M. E. J. (2006). A dynamical model of general intelligence:
The positive manifold of intelligence by mutualism. Psychological Review, 113, 842–861. https://doi.org/10.1037/0033-295X.113.4.842
Verdine, B. N., Irwin, C. M., Golinkoff, R. M., & Hirsh-Pasek, K. (2014). Contributions of executive function and spatial skills to preschool mathematics achievement.
Journal of Experimental Child Psychology, 126, 37–51. https://doi.org/10.1016/j.jecp.2014.02.012
Wechsler, D. (2003). Wechsler intelligence scale for children (4th ed.). The Psychological Corporation.
Winkelmann, H., Robitzsch, A., Stanat, P., & Köller, O. (2012). Mathematische Kompetenzen in der Grundschule. Diagnostica, 58, 15–30.

13
T. Kahl et al. Cognitive Development 62 (2022) 101149

Wynn, K. (1992). Addition and subtraction by human infants. Nature, 358(6389), 749–750. https://doi.org/10.1038/358749a0
Xie, F., Zhang, L., Chen, X., & Xin, Z. (2020). Is spatial ability related to mathematical ability: A meta-analysis. Educational Psychology Review, 32(1), 113–155. https://
doi.org/10.1007/s10648-019-09496-y
Yeniad, N., Malda, M., Mesman, J., van IJzendoorn, M. H., & Pieper, S. (2013). Shifting ability predicts math and reading performance in children: A meta-analytical
study. Learning and Individual Differences, 23, 1–9. https://doi.org/10.1016/j.lindif.2012.10.004

14