Psychological Research (2021) 85:36–46

https://doi.org/10.1007/s00426-019-01249-4

REVIEW

Executive Functions and Visual‑Spatial Skills Predict Mathematical

Achievement: Asymmetrical Associations Across Age
Tobias Kahl1 · Alexander Grob1 · Robin Segerer1 · Wenke Möhring1

Received: 20 March 2019 / Accepted: 14 September 2019 / Published online: 27 September 2019
© Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract
Children’s mathematical achievement depends on their domain-specific abilities and their domain-general skills such as
executive functions (EFs) and visual-spatial skills (VSS). Research indicates that these two domain-general skills predict
mathematical achievement. However, it is unclear whether these skills are differently associated with mathematical achieve-
ment across a large age range. The current cross-sectional study answered this question using a large, representative sample
aged 5–20 years (N = 1754). EFs, VSS, and mathematical achievement were assessed using the Intelligence and Develop-
ment Scales–2. Hierarchical regression analyses were computed with EFs and VSS as predictor variables and mathematical
achievement as dependent variable. We examined (non-) linear effects and interactions of EFs and VSS with age. Results
indicated that EFs and VSS were distinctly associated with mathematical achievement above and beyond effects of age, sex,
maternal education, and verbal reasoning. Effects of EFs were linear and age-invariant. Effects of VSS were curvilinear and
stronger in adolescents than in children. Our results indicated that EFs and VSS related differently to mathematical proficiency
across age, suggesting a varying impact on mathematics across age.

Introduction mathematical skills (e.g., Geary, 2000). On the other hand,

general cognitive processes play a key role for mathemati-
Mathematical achievement is essential for success in modern cal skills (LeFevre et al., 2010) with particularly executive
society, resulting in better job prospects and quality of life functions (EFs) and visual-spatial skills (VSS) predicting
(Geary, 2011a; Gross, Hudson, & Price, 2009). Although mathematical achievement (for overviews, see Cragg & Gil-
precursors such as early quantitative competencies emerge more, 2014; Mix & Cheng, 2012). The present study aims
long before entering school, mathematical skills are devel- at extending previous research by investigating the distinct
oped considerably in primary and secondary school (Geary, association of these two domain-general skills with math-
2000). However, which abilities and factors advance chil- ematical achievement using a large, representative sample
dren’s and adolescents’ mathematical skills? On the one even when accounting for the respective other skill and other
hand, research has indicated that children are predisposed to control variables. Importantly, we explore whether the rela-
represent and acquire numerical knowledge (e.g., Dehaene, tions of these two domain-general skills with mathematics
1992; Gelman, 1990; Wynn, 1992). Domain-specific abili- vary with age from preschool to young adulthood.
ties such as an understanding of numerosity or ordinality,
and the approximate number sense are present early in Executive functions and mathematics
life and provide the basis for subsequent, more complex
Miyake and Friedman (2012) define EFs as “a set of gen-
Electronic supplementary material The online version of this eral-purpose control processes that regulate one’s thoughts
article (https​://doi.org/10.1007/s0042​6-019-01249​-4) contains and behaviors” (p. 8). Several studies indicated that this
supplementary material, which is available to authorized users. construct consists of highly related but separable compo-
nents such as working memory, inhibition, and cognitive
* Tobias Kahl
[email protected] flexibility (e.g., Diamond, 2013; Lehto, Juujärvi, Koois-
tra, & Pulkkinen, 2003; Miyake et al., 2000). Working
1
Department of Psychology, Developmental and Personality memory enables to hold and manipulate information in
Psychology, University of Basel, Missionstrasse 62, mind (Baddeley, 1996). Several mathematical operations
4055 Basel, Switzerland

13
Vol:.(1234567890)
Psychological Research (2021) 85:36–46 37

require working memory, for example when a person keeps Visual‑spatial skills and mathematics
intermediate results in mind while continuing an opera-
tion. Research yielded strong evidence for a link between VSS are defined as the ability to generate, retrieve, and
working memory and mathematics in various age groups transform structured visual images (cf., Casey, Nuttal, &
(e.g., Cragg, Keeble, Richardson, Roome, & Gilmore, Benbow, 1995). Like EFs, VSS comprise different sub-
2017; for overviews, see Bull & Lee, 2014; Cragg & Gil- components, but researchers are discordant what the com-
more, 2014). Fewer studies have investigated inhibition ponents are (cf. Newcombe & Shipley, 2015). Research
and cognitive flexibility. Inhibition involves controlling revealed that different types of VSS relate to mathemati-
attention or behavior to suppress a dominant response cal achievement (e.g., Frick, 2018; Lauer & Lourenco,
(Diamond, 2013) and comes into play when incorrect, 2016; Verdine, Irwin, Golinkoff, & Hirsh-Pasek, 2014;
but more automatic number facts should be suppressed for reviews, see McCrink & Opfer, 2014; Mix & Cheng,
(e.g., inhibiting six when being asked to multiply three by 2012), with links found between mathematics and spatial
three). Several studies have shown that inhibitory skills assembly competencies (Kyttälä, Aunio, Lehto, Van Luit,
predict mathematical achievement (e.g., Bull & Scerif, & Hautamaki, 2003; Markey, 2010), mental rotation (Gun-
2001; Gilmore et al., 2013). Cognitive flexibility is based derson, Ramirez, Beilock, & Levine, 2012), and spatial
on working memory and inhibition and is defined as being scaling (Möhring, Newcombe, & Frick, 2015).
flexible and being able to change perspectives (Diamond, With respect to developmental changes in the magni-
2013). In mathematics, this ability is essential when chil- tude of these relations, it has often been proposed that VSS
dren have to quickly change strategies or switch between becomes increasingly associated with mathematics across
different types of tasks. In line with results on other com- development (Mix & Cheng, 2012). In line with this con-
ponents, it was found that cognitive flexibility predicted clusion, some studies have shown that relations between
mathematical achievement (for a meta-analysis, Yeniad, spatial skills and mathematics became indeed stronger
Malda, Mesman, van IJzendoorn, & Pieper, 2013). with increasing age (e.g., Li & Geary, 2013; Wolfgang,
Overall, several components of EFs seem to be strongly Stannard, & Jones, 2001). However, other studies showed
related to mathematical achievement at different ages. contrary results. For example, Holmes and Adams (2006)
However, so far, it is unclear whether (a) EFs as a unified found that spatial skills predicted unique variance in third
construct is related to mathematical achievement and (b) graders’ mathematics performance, but less variance in
whether the association of EFs with mathematical achieve- fifth graders’ performance. Again, other studies indicated
ment differs systematically across age. With respect to the stable relations between spatial and mathematical ability
first question, a recent study indicated no relation between in 4- to 11-year-olds (Hawes et al., 2018), 6- to 12-year-
mathematical achievement and EFs at the latent variable olds (Mix et al., 2016), and in 8- to 25-year-olds (Cragg
level (Hawes, Moss, Caswell, Seo, & Ansari, 2018) which et al., 2017). Overall, analogously to the link between EFs
seems surprising in light of studies about separate compo- and mathematics, it remains poorly understood whether
nents. With respect to age-related changes in this associa- space-math relations change or remain stable across a stu-
tion, results are heterogenous so far. Using a longitudinal dent’s career.
approach with children from first to fifth grade, Geary
(2011b) showed that the contribution of EFs to mathemat-
ics increased with age. Contrary to this result, another lon- The present study
gitudinal study (Stipek & Valentino, 2015) supported the
“fade-out” hypothesis, which proposes that EFs become The predictive power of EFs and VSS on mathematical
less important with age. This latter study indicated that EFs achievement has been repeatedly shown in several stud-
were vital for younger children’s mathematical achievement ies (Cragg & Gilmore, 2014; Mix & Cheng, 2012). How-
but less so for older children. A third possibility was sug- ever, so far, it remains unanswered whether these skills
gested by another recent study providing evidence for an are distinctly associated with mathematical achievement
age-invariant relation between EFs and mathematics (Cragg and whether these associations change across age. Prior
et al., 2017). More concretely, using several cross-sectional studies have revealed mixed results with respect to this lat-
samples between 8 and 25 years, the authors found stable ter question. Possible explanations for these mixed results
relations between mathematics and in particular working can be found in methodological differences among these
memory across the age groups tested. Overall, these mixed studies. Whereas some studies have used a longitudinal
results indicate that the question of how the association approach (e.g., Geary, 2011b), most of these studies used a
between EFs and mathematics changes across childhood, cross-sectional design (e.g., Cragg et al., 2017; Mix et al.,
adolescence, and young adulthood is insufficiently answered 2016). Furthermore, depending on the study, VSS, EFs,
as of today.

13
38 Psychological Research (2021) 85:36–46

and mathematical achievement have been differently oper- p < 0.05, η2 = 0.004, and lower scores on verbal reasoning,
ationalized and different control variables were included F(1, 1965) = 30.62, p < 0.001, η2 = 0.015, compared to par-
in the analyses. Moreover, these studies have often used ticipants who remained in the study.
varying age groups, with sometimes small age ranges that Participants were assessed in Switzerland (56.2%), Ger-
are not necessarily comparable (e.g., Hawes et al., 2018; many (39.2%), Austria, and Lichtenstein (4.7%). A total of
Verdine et al., 2014; for reviews, see Cragg & Gilmore, 76.0% were monolingual native German speakers, 15.3%
2014; Mix & Cheng, 2012). Importantly, no study to our reported being bilingual, and 8.5% reported another lan-
knowledge has analyzed effects of EFs and VSS concur- guage being their dominant language. However, participants’
rently on mathematics across a wide age range. Such an German level was sufficient to comprehend the task instruc-
endeavor may be crucial as working on a VSS task, such tions and to answer accordingly. The Ethics Committee of
as mentally manipulating spatial images, requires execu- Northwestern and Central Switzerland (EKNZ) approved the
tive resources. That is, controlling for EFs seems essential current study, which was performed in accordance with the
as it may clarify the predictive power of VSS per se on rules laid down in the 1964 Declaration of Helsinki and its
mathematical performance. later amendments. Participants and families provided con-
Building on this theoretical background, the aims of the sent for participation.
present study were two-fold. First, we investigated whether
EFs and VSS are distinctly associated with mathematical Measures
achievement. To this end, we used a large, cross-sectional
sample aged 5–20 years. Second, we explored whether asso- All variables of the present study were assessed with the
ciations of EFs and VSS with mathematical achievement same test battery, namely the IDS-2 (Grob & Hagmann-von
differ in strength across age. We controlled for sex, socioeco- Arx, 2018). The IDS-2 is an extensive test battery that ena-
nomic status, and verbal reasoning in the statistical analy- bles to measure a wide range of developmentally relevant
ses given that these variables have been shown to relate to competencies of children, adolescents, and young adults.
the variables of interest and such relations were beyond the More concrete, this instrument allows to measure partici-
scope of the present study (e.g., mathematical performance: pants’ intelligence, executive functions, psychomotor skills,
Baird, 2012; Hoff, 2013; Lindberg, Hyde, Petersen, & Linn, social-emotional competencies, scholastic skills, and atti-
2010; for a meta-analysis: Kuncel, Hezlett, & Ones, 2004; tudes towards work. However, for the present study, only
Peng, Wang, Wang, & Lin, 2018; spatial ability: Levine, particular subtests of the constructs intelligence, executive
Foley, Lourenco, Ehrlich, & Ratliff, 2016; Levine, Hut- functions, and scholastic skills were included (for details, see
tenlocher, Taylor, & Langrock, 1999; Levine, Vasilyeva, below). The IDS-2 is a revised version of the Intelligence
Lourenco, Newcombe, & Huttenlocher, 2005). and Development Scales (IDS, Grob, Meyer, & Hagmann-
von Arx, 2009) and is currently validated in a number of
studies (for initial evidence on high construct validity, cf.
Method Grieder & Grob, 2019). The previous version of the IDS-2
was shown to be highly reliable and valid (Hagmann-von
Participants Arx, Grob, Petermann, & Daseking, 2012; Hagmann-von
Arx, Meyer, & Grob, 2008; Hagmann-von Arx, Petermann,
Participants (N = 2030, 51.8% female) were drawn from the & Grob, 2013).
standardization sample of the Intelligence and Develop- Testing took place either at the laboratory of the respec-
ment Scales-2 (IDS-2, Grob & Hagmann-von Arx, 2018). tive University or at participants’ homes. As is customary in
Participants with missing data in the variables EFs, VSS, intelligence testing, children started at different points of the
mathematical achievement, or in the control variables were test depending on their age and finished the test depending
excluded (n = 236). Furthermore, we excluded participants on their performance.
with z-scores beyond ± 2.58 on EFs, VSS, or mathematical
achievement as these participants scored beyond 99% of the Mathematical achievement
remaining sample and their scores may bias the statistical
procedures (cf. Field, 2013, n = 40). This resulted in a final This task measured a broad range of mathematical achieve-
sample of 1754 participants (52.5% female) aged 5–20 years ment using 19 sections, with each section consisting of
(M = 12.33 years, SD = 4.44). The distribution of the sample several items. Content was based on the Swiss school cur-
across age can be seen in Supplementary Table 1. To see riculum (Erziehungsdirektoren-Konferenz, D-EDK, 2013).
whether excluded participants differed on their SES or ver- Five- to 7-year-olds started with tasks assessing mathe-
bal reasoning skills, a MANOVA was conducted. Excluded matical precursor skills, for example counting, magnitude
participants showed a lower SES, F(1, 1965) = 8.11, understanding, or knowledge about invariance (e.g., children

13
Psychological Research (2021) 85:36–46 39

Fig. 1  A selection of three tasks (© Hogrefe AG, Bern: printed with sional geometrical figures and asked to copy them using triangular
kind permission of Hogrefe AG, Bern) used to measure inhibition (a) or rectangular cutouts. In the second assembly task (c), participants
and visual-spatial skills (b, c). In the inhibition task (a), participants were presented with a printed pattern of circles and were asked to
were asked to name the correct color of each presented animal. In the copy this pattern using thin plates
first assembly task (b), participants were presented with 2-dimen-

were asked to count small wooden cubes in accordance to compared to a control group with average competencies in
a prior presented number). Eight- to 10-year-olds started mathematics (Grob & Hagmann-von Arx, 2018), which sug-
with problems on equations, proportions, or mental addi- gests a high differential validity of this subtest.
tion (e.g., “There are 15 dices in a box. If I would add
three—how many dices would be in the box?”). Participants Executive functions
aged 11 years and older started with items measuring their
knowledge about advanced proportional reasoning, geom- Three tasks measuring participants’ inhibition, cognitive
etry, algebra, or fractions (e.g., participants were asked to flexibility, and updating skills were used. The inhibition task
compare the magnitude of fractions and decimal numbers). was an adapted Stroop task (Stroop, 1935) in which partici-
Based on extensive pilot studies and as is common in other pants were asked to name the correct color of four different
test batteries (e.g. Wechsler Intelligence Scale for Children, animals (dolphin, chick, frog, ladybug). Similar Stroop tasks
WISC-IV, Wechsler, 2003), participants were assumed to with for instance fruits have been used in other recent stud-
solve the items before their starting point correctly. Diffi- ies (cf. Neuenschwander, Röthlisberger, Cimeli, & Roebers,
culty increased within every section. The task was stopped 2012; Röthlisberger, Neuenschwander, Cimeli, & Roebers,
when participants answered incorrectly in five subsequent 2013). The task consisted of three conditions. In the first
items. The first eight sections were untimed whereas items condition, children were asked to name the correct color of
of the following sections had a time limit of each 90s. The each animal in a set of 36 congruently colored animals as
number of correct answers served as dependent variable. quickly and accurately as possible. In the second condition,
This subtest showed a high reliability (Cronbach’s α = 0.97). participants were asked to name the correct color of each
External validity of this subtest was indicated by significant animal in a set of 36 black-white presented animals. In the
correlations with student’s math grades (r = 0.37, N = 478) final condition, they were asked to name the correct color of
and their mathematical competencies based on parents’ each animal in a set of 36 incongruently colored animals (see
report (r = 0.44, N = 726; Grob & Hagmann-von Arx, 2018). Fig. 1a). The times of completion for each condition (t1, t2,
Furthermore, it was found that mathematically gifted stu- and t3) were measured. Following previous work in this area
dents performed significantly better in this math test as (Archibald & Kerns, 1999; Röthlisberger, Neuenschwander,

13
40 Psychological Research (2021) 85:36–46

Michel, & Roebers, 2010) an inhibition score was created performance ranging between zero (i.e., no overlap with the
using the following formula (t3 − [t1 × t2]/[t1 + t2]). The for- target) to two points (i.e., exact overlap) for every circle.
mula represents a measure of interference with lower values The number of points served as dependent variable. These
representing higher inhibition skills. assembly tests showed high reliabilities (Cronbach’s αs:
To assess cognitive flexibility, participants had to name as 0.92–0.95)
many examples as possible in a given category within 90s.
Such verbal fluency tasks have often been used as a measure Verbal reasoning
of children’s cognitive flexibility in previous studies (van der
Elst, Hurks, Wassenberg, Meijs, & Jolles, 2011; for a sum- Verbal reasoning was assessed using two tasks, each consist-
mary, cf. Diamond, 2013). Five- to 9-year-olds were asked to ing of 34 items. In one task, participants were asked to name
name examples within the categories “animals” and “food”. the opposite of words (e.g., poor). In the other task, par-
Ten- to 20-year-olds were additionally presented with the ticipants were presented with a set of three pictures of one
categories “words beginning with E” and “words beginning category (i.e., showing different cars) and asked to name the
with alternatingly S or L”. The number of correct words category. Difficulty increased and the task was stopped when
across target categories served as cognitive flexibility score. participants gave three subsequent false answers. Neither
To measure updating, participants had to recall differ- task used a time limit. In both tasks, the number of correct
ent letters in a backward order as has been typically done answers served as dependent variables. Reliabilities for both
in other intelligence tests (e.g., WISC-IV, Wechsler, 2003). tasks were high (Cronbach’s αs: 0.91–0.94).
The task consisted of 10 trials. Participants started with
two letters and one additional letter was added stepwise to Socioeconomic status
increase difficulty. Participants were presented with a maxi-
mum of nine letters. The task was stopped when partici- SES was estimated based on the mother’s educational level
pants produced three subsequent false recalls. The number as assessed in parent questionnaires (cf. Hoff, 2013). These
of correctly recalled series was used as an updating score. educational levels ranged from obligatory school education
The subtests measuring EFs showed satisfying reliabilities to university degrees (on a five-point Likert scale).
(Cronbach’s αs: 0.71–0.75; Schmitt, 1996).
Data analytic procedure

Visual‑spatial skills Cross-sectional data were analyzed using IBM SPSS Statis-
tics 23.0. To investigate whether EFs and VSS independently
Two different assembly tasks measuring VSS were used. predicted mathematical achievement after accounting for age,
Such assembly tasks have often been used to operationalize sex, SES, and verbal reasoning, a hierarchical linear regres-
spatial skills (e.g., in Jirout & Newcombe, 2015; Kyttäla sion analysis was conducted. To examine whether associations
et al., 2003; Markey, 2010). In the first task, participants of EFs and VSS with mathematical achievement changed
were asked to replicate two-dimensional figures by using across the different age groups, interaction terms involving
triangular or rectangular cutouts (see Fig. 1b). The task con- age (e.g., age*VSS) were entered into the model. As relations
sisted of 20 items. The figures became more complex by of mathematical achievement with age, EFs, or VSS may fol-
either increasing the number of the required cutouts or by low non-linear patterns (cf. age: Geary, 2000; EFs: Stipek &
blurring the edges of the required cutouts in the to-be-copied Valentino, 2015; VSS: Li & Geary, 2013), quadratic terms
two-dimensional pattern (similar to the block design subtest of age, EFs, and VSS were also entered into the model. By
in the WISC-IV). The task was stopped when participants including these terms, we were able to distinguish between
produced three subsequent false replications. Time to assem- quadratic main effects (e.g., VSS*VSS) and interaction effects
ble the configuration was limited for each item (ranging from between the variables (Darlington & Hayes, 2016). Significant
30 to 120 s). The number of correctly replicated configura- interactions were analyzed and visualized by simple slopes
tions was used as dependent variable. analyses (cf. Aiken, West, & Reno, 1991).
Similarly, in the second assembly task, participants were
asked to reproduce a pattern (see Fig. 1c). They were pre-
sented with a printed pattern of circles and asked to copy Results
this pattern using thin plates on a separate sheet. Partici-
pants started with the first pattern using a time limit of 60s. Preliminary analyses
The task consisted of four trials with increasing complex-
ity. Accuracy was assessed by comparing the produced pat- As the constructs under investigation (EFs, VSS, and ver-
tern to a stencil. Participants received points for accurate bal reasoning) were operationalized with several tasks,

13
Psychological Research (2021) 85:36–46 41

Table 1  Partial correlations Variable 1 2 3 4 5 6

between executive functions,
visual-spatial skills, 1. Sex _
mathematical achievement, and
2. SES 0.03 _
the control variables sex, SES,
and verbal reasoning (n = 1754) 3. Verbal reasoning 0.03 0.22** _
after accounting for age 4. EFs 0.12* 0.17** 0.54** _
5. VSS − 0.13 0.16** 0.46** 0.47** _
6. Mathematical achievement − 0.10* 0.17** 0.58** 0.59** 0.57** _

Factor scores were used for EFs, VSS, and verbal reasoning. EFs = executive functions; VSS = visual-spa-
tial skills; SES = socioeconomic status; Sex: 0 = male, 1 = female
*p < 0.05, **p < 0.001

Table 2  Relations Among Variables B β T Sign. T F Sign. F

R2
EFs, VSS, and Mathematical (df1, df2)
Achievement: Hierarchical
Regression Analysis Step 1 Age 0.973 0.316 14.911 < 0.001 0.79 1677 < 0.001
Sex − 1.725 − 0.063 − 5.773 < 0.001 (4, 1749)
SES 0.252 0.029 2.571 0.010
Verbal reasoning 8.327 0.608 28.740 < 0.001
Step 2 EFs 4.465 0.325 16.236 < 0.001 0.85 1642 < 0.001
VSS 3.716 0.271 13.688 < 0.001 (6, 1747)
Step 3 EFs*Age 0.143 0.043 1.344 0.179 0.86 964 < 0.001
VSS*Age 0.255 0.070 2.441 0.015 (11, 1742)
EFs*EFs − 0.033 − 0.003 − 0.131 0.896
VSS*VSS − 0.893 − 0.066 − 3.543 < 0.001
Age*Age − 0.122 − 0.170 − 7.100 < 0.001

EFs executive functions, VSS visual-spatial skills, SES socioeconomic status, Age mean-centered age, Sex:
0 = male, 1 = female

we aggregated these indicators to obtain a single value for reached by a VIF value beyond 10 and a tolerance value
each construct (see Supplementary Table 2 for descriptive below 0.10 (e.g., O’brien, 2007).
statistics of each single task and Supplementary Table 3
for bivariate and partial correlations between the separate Relations of EFs and VSS with mathematical
tasks). To obtain single scores for these constructs, we esti- achievement across age
mated principle-axis regression factor scores, each based on
a one-factor solution (Di Stefano, Zhu, & Mîndrila, 2009). The regression model revealed distinct relations between EFs
These extracted factor scores are z-standardized and com- and VSS with mathematical achievement above and beyond
prise the shared variance of the tasks. Thus, factor scores effects of age, sex, SES, and verbal reasoning (see Table 2).
are considered to be a measure of error-free estimates of the That is, the higher participants scored on tasks assessing EFs
latent abilities of individuals. and VSS, the higher they scored in the mathematical task.
It was found that these factor scores accounted for large The initial model at step 1 included the control variables
portions of variance in the indicators of VSS (87.94%), (age, sex, SES, and verbal reasoning) and explained 79%
EFs (76.92%), and verbal reasoning (94.35%), thus reflect- of the variance in mathematical achievement. Adding EFs
ing large homogeneity across the tasks used. Factor scores and VSS in a second step, increased the explained variance
and the other variables under investigation were moder- significantly (6%) as did the interaction and quadratic effect
ately correlated with each other, when controlling for age terms (1%) in a third step. With respect to the interaction
(see Table 1). Multicollinearity analyses at the factor score and quadratic effects, it was found that the quadratic terms of
level showed acceptable tolerance and VIF statistics (EFs: VSS were significant and moderated by age. Effects of EFs
tolerance = 0.214, VIF = 4.664: VSS: tolerance = 0.221, were neither quadratic nor dependent on age. As the interac-
VIF = 4.533). A concerning level of multicollinearity is tion terms EF*EF*age and VSS*VSS*age did not explain

13
42 Psychological Research (2021) 85:36–46

(A) 50 mathematical task, which remained similar for children as

well as adolescents.
40 Given that the current test of mathematical achievement
Mathematical Achievement

was based on the curriculum, it was the case that older chil-
dren were more likely asked to solve geometry items as com-
(Raw Scores)

30
pared to younger children. Therefore, it is possible that the
Children above-mentioned effects indicating that VSS and mathemat-
20
Adolescents
ics are more strongly related in adolescents as compared to
10
children reflect adolescents’ increased likelihood of solving
geometry items. To examine this possibility, we computed
an additional analysis. In a first step, two independent raters
0
-3 -2 -1 0 1 2 3 categorized items of the mathematics task as assessing
Visual Spatial Skills (Z-Scores) geometry or not (showing a high inter-rater reliability of
IRR = 1). Eight out of 64 items were rated to assess geom-
(B) 50
etry knowledge. In a next step, all analyses described above
were conducted again with a new mathematics score exclud-
40
Mathematical Achievement

ing these eight items assessing geometry knowledge. Results

showed identical effects (see Table 3).1 Even though the beta
(Raw Scores)

30 coefficients of the interaction VSS*age became smaller as

Children compared to the analysis with the full mathematics score
20 (βwithout geometry = 0.059 vs. βwith geometry = 0.070), the interac-
Adolescents
tion was still significant. Therefore, it seems that—even after
10 excluding geometry items—the association between VSS
and mathematics seems stronger in adolescents as compared
0 to children.
-3 -2 -1 0 1 2 3
Executive Function (Z-Scores)

Discussion
Fig. 2  Participants’ mathematical achievement scores as a function
of individual differences of their visual-spatial skills (a) or executive
functions (b). Separate curves are shown for participants with one The current study investigated relations among EFs, VSS,
standard deviation above (i.e., adolescents) or below the mean age and mathematical achievement in children, adolescents, and
(i.e., children) young adults, and assessed whether associations between
these variables changed in quality across this age range.
any further variance, these terms were excluded from the Our findings indicate that EFs and VSS predict mathematics
model. independently, which is in line with several studies showing
Follow-up single slope analyses (Aiken et al., 1991) similar predictions in various age groups (for overviews, see
were conducted to reveal the nature of the interaction effect Cragg & Gilmore, 2014; Mix & Cheng, 2012). Our results
VSS*age and to illustrate this effect graphically. As is com- also extend and qualify these previous findings. Using a
mon in research investigating interaction effects, the relation large, representative sample aged 5–20 years, our findings
between VSS and mathematical achievement is displayed indicate for the first time that relations between EFs and
for one standard deviation above (i.e., adolescents) or below mathematical achievement are age-invariant whereas effects
the mean age in our sample (i.e., children; see Fig. 2a). The
full regression models for the estimated effects of VSS on
mathematical achievement for children and adolescents are 1
According to the instructions, the task was stopped when partici-
shown in Supplementary Table 4 and 5, respectively. As can pants answered incorrectly in five subsequent items. Consequently,
be seen in Fig. 2a, associations between VSS and mathemati- in some subjects, the termination of the test could also have been
cal achievement are stronger in adolescents than in children. caused by geometry items. In order to achieve a complete correc-
tion of the test results with regard to the contributions of the geom-
However, the quadratic effect shows that with increasing etry items, we estimated latent mathematical ability scores based on
scores on VSS tasks, the relation with mathematical achieve- a two-parametric item-response model with the data of the completed
ment becomes smaller. Contrary to VSS, the effect of EFs math items only. We excluded all geometry items from the model. In
on mathematical achievement was linear and unconditional this latent variable approach, the interaction term VSS*age showed
a tendency (p = 0.057). The size of the effect, however, appears only
on age. As can be seen in Fig. 2b, the higher participants slightly reduced as compared to the analyses with geometry items
scored on tasks assessing EFs, the higher they scored in the (βwithout geometry = 0.057 vs. βwith geometry = 0.070).

13
Psychological Research (2021) 85:36–46 43

Table 3  Relations Among Variables B β T Sign. T F Sign. F

R2
EFs, VSS, and Mathematical (df1, df2)
Achievement with Excluded
Geometry Items: Hierarchical Step 1 Age 0.873 0.312 14.477 < 0.001 0.79 1597 < 0.001
Regression Analysis (4, 1749)
Sex − 1.335 − 0.054 − 4.831 < 0.001
SES 0.191 0.024 2.112 0.035
Verbal reasoning 7.535 0.606 28.131 < 0.001
Step 2 EFs 4.419 0.355 17.712 < 0.001 0.85 1639 < 0.001
VSS 3.488 0.281 14.161 < 0.001 (6, 1747)
Step 3 EFs*Age 0.080 0.026 0.857 0.391 0.87 1057 < 0.001
VSS*Age 0.195 0.059 2.144 0.032 (11, 1742)
EFs*EFs − 0.195 − 0.018 − 0.894 0.371
VSS*VSS − 1.106 − 0.090 − 5.035 < 0.001
Age*Age − 0.115 − 0.177 − 7.694 < 0.001

EFs executive functions, VSS visual-spatial skills, SES socioeconomic status, Age mean-centered age, Sex:
0 = male, 1 = female

of VSS on mathematics are age-dependent and more pro- in our sample’s mathematical thinking. This goes in line
nounced for adolescents than for children. with results shown by Verdine et al. (2014), in which age,
Research has repeatedly shown that EFs and VSS are vocabulary knowledge, and SES explained already 61% of
associated with mathematical achievement (for reviews, the variance in a sample of 44 three- to four-year-olds. By
see Cragg & Gilmore, 2014; Mix & Cheng, 2012). How- contrast to this latter study, effects of the present study were
ever, the respective studies mainly focused on specific time shown for the entire period of children’s school career. Our
points in development and have often used small sample findings provide valuable insights for students’ mathematical
sizes. Furthermore, previous studies rarely examined VSS development. This information is central to progress in psy-
and EFs simultaneously and thus, accounted for the influ- chological theory, indicating that EFs and VSS need to be
ence of the respective other skill. Moreover, there is some considered in theoretical models that define the importance
heterogeneity with respect to how authors define constructs of domain-general skills on mathematical achievement. Fur-
such as visual-spatial working memory (VSWM). Whereas thermore, our results are essential for training and measuring
some authors interpreted outcomes of this measure as indi- mathematical achievement considering that EFs and VSS are
cating executive functioning (e.g. Cragg et al., 2017; Van often neglected in intervention programs and diagnostics.
de Weijer-Bergsma, Kroesbergen, & Van Luit, 2015), oth- Few studies investigated whether these associations of
ers used VSWM as a marker for spatial skills (for an over- EFs and VSS on mathematics change across age and results
view, cf. Mix & Cheng, 2012). A recent study from Hawes to date are heterogeneous. That is, for EFs (e.g., Cragg et al.,
et al. (2018) indicated that VSWM is better characterized as 2017; Geary, 2011b; Stipek & Valentino, 2015) and for VSS
indicating EFs; however, more research is needed to more (e.g., Cragg et al., 2017; Hawes et al., 1018; Mix et al., 2016;
clearly differentiate spatial measures from those indicating Wolfgang et al., 2001), there is no clear evidence whether
executive functioning. the magnitudes of relations to mathematics change or remain
Our study allowed to address those previous methodo- stable across age. The current findings demonstrate that rela-
logical considerations as our large, cross-sectional sample tions between EFs and mathematics are age-invariant (for
spanned development from 5 to 20 years. Furthermore, we similar results, see Cragg et al., 2017). Thus, it seems that
used reliable subtests and refrained from including VSWM. EFs are continuously important across childhood, adoles-
In the statistical analyses, we entered EFs and VSS simul- cence, and young adulthood. Contrary to that finding, the
taneously while accounting for a number of control vari- association of VSS with mathematical achievement was age-
ables in order to reveal the relative contribution of each dependent with stronger relations of VSS and mathematics
domain-general skill to mathematical achievement. Our for adolescents as compared to children. One reason for this
results revealed that EFs and VSS are significantly asso- finding could be that adolescents are more likely solving
ciated with mathematical achievement, even after other geometry items as compared to younger participants as they
strong predictors such as verbal reasoning were accounted advance to higher mathematics. To investigate this expla-
for. Especially age and verbal reasoning correlated highly nation, we conducted an additional analysis and excluded
with mathematical achievement and together with other con- geometry items. It was found that even after removing
trol variables explained approximately 80% of the variance these geometry items from the mathematics score, effects

13
44 Psychological Research (2021) 85:36–46

remained similar. The beta coefficients of the VSS*age In addition, it needs to be noted that our design was cross-
interaction decreased slightly indicating that geometry items sectional, precluding conclusions about the causal effects of
did have an impact on the results. However, the interaction EFs and VSS on mathematics. Future studies with longitudi-
between VSS and age was still significant, highlighting that nal designs and training studies should investigate potential
this interaction does not simply persist because adolescents causal relations.
solve more geometry items. Overall, our results indicate that Using a large, representative sample, the current study
spatial ability seems particularly crucial when learning and provides evidence that EFs and VSS are distinctly related to
coping with more complex mathematics as opposed to basic mathematical achievement. Our study widens our in-depth
number concepts or operations (cf. Mix & Cheng, 2012). understanding of domain-general skills and their link to
Beyond clarifying the associations of EFs and VSS with mathematical achievement by clarifying the age-invariant
mathematics, the current study also investigated whether or age-specific association across a wide age range.
these associations across age follow a linear or a non-linear
pattern. Effects of EFs on mathematical achievement fol- Acknowledgements We are grateful to Priska Hagmann-von Arx and
Nora Newcombe for their input on the present research questions. Fur-
lowed a linear pattern across age, implying that higher EFs ther, we thank our colleagues of the Division of Developmental and
are constantly related to higher mathematical achievement. Personality Psychology for their helpful feedback during the brown bag
On the contrary, effects of VSS on mathematics followed meetings. A special thank goes to the research assistants who were in
a curvilinear pattern. Our data suggested that the relation charge of data collection.
between VSS and mathematics did not follow a proportional
association with “the more [VSS] the better” for mathemati- Compliance with ethical standards
cal achievement. Instead, it seemed that this effect reached
Conflict of interest The authors declare that they had no conflicts of in-
a ceiling at a certain high level of VSS at which point the terest with respect to their authorship or the publication of this article.
association of VSS with mathematics became small to non-
existent. These results bear important implications for inter-
vention programs and school curricula. One conclusion with
respect to EFs may be that even improving EFs on a high References
level may make a difference on mathematical achievement.
With respect to VSS, results implicate that children and ado- Aiken, L. S., West, S. G., & Reno, R. R. (1991). Multiple regression:
lescents with poor VSS may profit more from training VSS testing and interpreting interactions. London: SAGE. https​://doi.
than participants with high VSS. org/10.1037/10520​-147.
Archibald, S. J., & Kerns, K. A. (1999). Identification and description
In the current study, we consider it an advantage that par- of new tests of executive functioning in children. Child Neuropsy-
ticipants were tested with the same standardized instrument, chology, 5, 115–129.
covering every construct under investigation in a conceptu- Baddeley, A. (1996). Exploring the central executive. The Quarterly
ally equivalent manner. In case of the mathematical assess- Journal of Experimental Psychology Section, 49A(1), 5–28. https​
://doi.org/10.1080/71375​5608.
ment tool, the task was age-specific and based on the school Baird, K. (2012). Class in the classroom: the relationship between
curriculum. Thus, this measure covers what students are school resources and math performance among low socioeco-
confronted with at school and items measured performance nomic status students in 19 rich countries. Education Economics,
in a variety of mathematical topics and operations. Conse- 20(5), 484–509.
Bull, R., & Lee, K. (2014). Executive functioning and mathematics
quently, our study shows proximity to how mathematical achievement. Child Development Perspectives, 8(1), 36–41. https​
achievement is defined in a school context and thus, bears ://doi.org/10.1111/cdep.12059​.
high ecological validity. Bull, R., & Scerif, G. (2001). Executive functioning as a predictor of
At the same time, our study shows some limitations. The children’s mathematics ability: inhibition, switching, and working
memory. Developmental Psychology, 19(3), 273–293. https​://doi.
mathematical achievement test was developed to measure org/10.1207/S1532​6942D​N1903​_3.
mathematical knowledge on topics of a student’s entire Casey, M., Nuttall, R., & Benbow, C. (1995). The influence of spatial
school career between kindergarten until high school. Con- ability on gender differences in mathematics college entrance test-
sequently, results do not provide information about the rela- scores across diverse samples. Developmental Psychology, 31(4),
697–705. https​://doi.org/10.1037/0012-1649.31.4.697.
tions between specific mathematical topics such as arithme- Cragg, L., & Gilmore, C. (2014). Skills underlying mathematics: The
tic, geometry, or word problems and EFs and VSS that hold role of executive function in the development of mathematics pro-
for the entire sample. Future studies may focus on isolated ficiency. Trends in Neuroscience and Education, 3(2), 63–68. https​
mathematical skills and their specific associations to EFs ://doi.org/10.1016/j.tine.2013.12.001.
Cragg, L., Keeble, S., Richardson, S., Roome, H. E., & Gilmore, C.
and VSS across age (as for example in Cragg et al., 2017). (2017). Direct and indirect influences of executive functions on
Furthermore, our pattern of relations between EFs, VSS, and mathematics achievement. Cognition, 162, 12–26. https​://doi.
math is restricted to one specific type of spatial skills and org/10.1016/j.cogni​tion.2017.01.014.
may differ with other spatial abilities (cf. Mix et al., 2016).

13
Psychological Research (2021) 85:36–46 45

Darlington, R. B., & Hayes, A. F. (2016). Regression analysis and Migrationshintergrund. Diagnostica, 59, 170–182. https​://doi.
linear models: Concepts, applications, and implementation. New org/10.1026/0012-1924/a0000​91.
York: Guilford Publications. Hawes, Z., Moss, J., Caswell, B., Seo, J., & Ansari, D. (2018). Rela-
Dehaene, S. (1992). Varieties of numerical abilities. Cognition, 44(1– tions between numerical, spatial, and executive function skills and
2), 1–42. https​://doi.org/10.1016/0010-0277(92)90049​-N. mathematics achievement: A latent-variable approach. Cognitive
Di Stefano, C., Zhu, M., & Mîndrilă, D. (2009). Understanding and Psychology, 109, 1–23.
using factor scores: Considerations for the applied researcher. Hoff, E. (2013). Interpreting the early language trajectories of chil-
Practical Assessment, Research & Evaluation, 14(20), 1–11. dren from low SES and language minority homes: Implications
Diamond, A. (2013). Executive functions. Annual Review of Psychol- for closing achievement gaps. Developmental Psychology, 49(1),
ogy, 64, 135–168. https​://doi.org/10.1146/annur​ev-psych​-11301​ 4–14. https​://doi.org/10.1037/a0027​238.
1-14375​0. Holmes, J., & Adams, J. W. (2006). Working Memory and Children’s
Erziehungsdirektoren-Konferenz (D-EDK) (2013). Lehrplan 21, Math- Mathematical Skills: Implications for mathematical develop-
ematik (Konsultationsfassung). Luzern: D-EDK ment and mathematics curricula. Educational Psychology, 26(3),
Field, A. (2013). Discovering Statistics Using SPSS (4th ed.). London: 339–366.
SAGE. Jirout, J. J., & Newcombe, N. S. (2015). Building blocks for devel-
Frick, A. (2018). Spatial transformation abilities and their relation oping spatial skills: Evidence from a large, Representative U.S.
to later mathematics performance. Psychological Research. Sample. Psychological Science, 26(3), 302–310. https​://doi.
https​://doi.org/10.1007/s0042​6-018-1008-5. (Advance online org/10.1177/09567​97614​56333​8.
publication). Kuncel, N. R., Hezlett, S. A., & Ones, D. S. (2004). Academic per-
Geary, D. C. (2000). From infancy to adulthood: The development of formance, career potential, creativity, and job performance: Can
numerical abilities. European Child and Adolescent Psychiatry, one construct predict them all? Journal of Personality and Social
9, 11–16. https​://doi.org/10.1007/s0078​70070​004. Psychology, 86(1), 148–161.
Geary, D. C. (2011a). Consequences, characteristics, and causes of Kyttälä, M., Aunio, P., Lehto, J. E., Van Luit, J., & Hautamaki, J.
mathematical learning disabilities and persistent low achieve- (2003). Visuospatial working memory and early numeracy. Edu-
ment in mathematics. Journal of Developmental and Behavioral cational and Child Psychology, 20, 65–76.
Pediatrics, 32(3), 250–263. https​://doi.org/10.1097/DBP.0b013​ Lauer, J. E., & Lourenco, S. F. (2016). Spatial processing in infancy
e3182​09ede​f. predicts both spatial and mathematical aptitude in child-
Geary, D. C. (2011b). Cognitive predictors of achievement growth hood. Psychological Science, 27(10), 1291–1298. https​://doi.
in mathematics: A five year longitudinal study. Developmental org/10.1177/09567​97616​65597​7.
Psychology, 47(6), 1539–1552. https​://doi.org/10.1037/a0025​ LeFevre, J.-A., Fast, L., Skwarchuk, S.-L., Smith-Chant, B. L.,
510. Bisanz, J., Kamawar, D., & Penner-Wilger, M. (2010). Pathways
Gelman, R. (1990). First principles organize attention to and learning to mathematics: Longitudinal predictors of performance. Child
about relevant data: Number and the animate-inanimate distinc- Development, 81(6), 1753–1767.
tion as examples. Cognitive Science, 14, 79–106. https​://doi. Lehto, J. E., Juujärvi, P., Kooistra, L., & Pulkkinen, L. (2003).
org/10.1207/s1551​6709c​og140​1_5. Dimensions of executive functioning: Evidence from children.
Gilmore, C., Attridge, N., Clayton, S., Cragg, L., Johnson, S., Mar- British Journal of Developmental Psychology, 21, 59–80.
low, N., & Inglis, M. (2013). Individual differences in inhibi- Levine, S. C., Dulaney, A., Lourenco, S. F., Ehrlich, S., & Ratliff,
tory control, not non-verbal number acuity, correlate with K. (2016). Sex differences in spatial cognition: Advancing the
mathematics achievement. PLOS One, 8(6), e67374. https​: // conversation. WIREs Cognitive Science, 7, 127–155.
doi.org/10.1371/journ​al.pone.00673​74. Levine, S. C., Huttenlocher, J., Taylor, A., & Langrock, A. (1999).
Grieder, S., & Grob, A. (2019). Exploratory factor analysis of the Early sex differences in spatial skill. Developmental Psychol-
intelligence and development scales–2: Implications for theory ogy, 35, 940–949.
and practice. Assessment., 1, 1. https​://doi.org/10.1177/10731​ Levine, S. C., Vasilyeva, M., Lourenco, S., Newcombe, N., & Hut-
91119​84505​1. (Advance online publication). tenlocher, J. (2005). Socioeconomic status modifies the sex dif-
Grob, A., & Hagmann-von Arx, P. (2018). Intelligence and develop- ference in spatial skill. Psychological Science, 16, 841–845.
mental Scales—2 (IDS-2). Bern: Hogrefe. Li, Y., & Geary, D. C. (2013). Developmental gains in visuospatial
Grob, A., Meyer, C. S., & Hagmann-von Arx, P. (2009). Intelligence memory predict gains in mathematics achievement. PLoS One,
and Development Scales (IDS). Bern: Huber. 8(7), e70160. https​://doi.org/10.1371/journ​al.pone.00701​60.
Gross, J., Hudson, C., & Price, D. (2009). The Long Term Costs of Lindberg, S. M., Hyde, J. S., Petersen, J. L., & Linn, M. C. (2010).
Numeracy Difficulties. London: Every Child a Chance Trust. New trends in gender and mathematics performance: A meta-
Gunderson, E. A., Ramirez, G., Beilock, S. L., & Levine, S. C. analysis. Psychological Bulletin, 136(6), 1123–1135. https​://
(2012). The relation between spatial skill and early number doi.org/10.1037/a0021​276.
knowledge: The role of the linear number line. Developmental Markey, S. M. (2010). The relationship between visual-spatial rea-
Psychology, 48(5), 1229–1241. https​://doi.org/10.1037/a0027​ soning ability and math and geometry problem solving. Dis-
433. sertation Abstracts International: Section B: The Sciences and
Hagmann-von Arx, P., Grob, A., Petermann, F., & Daseking, M. Engineering, 70, 7874.
(2012). Konkurrente Validität des HAWIK-IV und der Intelli- McCrink, K., & Opfer, J. E. (2014). Development of spatial-numer-
gence and Development Scales (IDS). Zeitschrift für Kinder- und ical associations. Psychological Science, 23(6), 439–445. https​
Jugendpsychiatrie und Psychotherapie, 40, 41–50. https​://doi. ://doi.org/10.1177/09637​21414​54975​1.
org/10.1024/1422-4917/a0001​48. Mix, K. S., & Cheng, Y. L. (2012). The relation between space and
Hagmann-von Arx, P., Meyer, C. S., & Grob, A. (2008). Assessing math: developmental and educational implications. In J. B. Ben-
intellectual giftedness with the WISC-IV and the IDS. Journal of son (Ed.), Advances in Child Development and Behavior (Vol.
Psychology, 216, 173–180. 42, pp. 197–243). San Diego: Elsevier Academic Press Inc.
Hagmann-von Arx, P., Petermann, F., & Grob, A. (2013). Kon- https​://doi.org/10.1016/b978-0-12-39438​8-0.00006​-x.
vergente und diskriminante Validität der WISC-IV und der Mix, K. S., Levine, S. C., Cheng, Y.-L., Young, C., Hambrick, D. Z.,
Intelligence and Development Scales (IDS) bei Kindern mit Ping, R., & Konstantopoulos, S. (2016). The latent structure of

13
46 Psychological Research (2021) 85:36–46

spatial skills and mathematics across development: Highly cor- Stipek, D., & Valentino, R. A. (2015). Early childhood memory and
related but separate domains. JEP General, 145(9), 1206–1227. attention as predictors of academic growth trajectories. Jour-
Miyake, A., & Friedman, N. P. (2012). The nature and organiza- nal of Educational Psychology, 107(3), 771–788. https​://doi.
tion of individual differences in executive functions. Current org/10.1037/edu00​00004​.
Directions in Psychological Science, 21(1), 8–14. https​://doi. Stroop, J. R. (1935). Studies of interference in serial verbal reactions.
org/10.1006/cogp.1999.0734. Journal of Experimental Psychology, 18, 643–662. https​://doi.
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., How- org/10.1037/h0054​651.
erter, A., & Wager, T. D. (2000). The unity and diversity of Van de Weijer-Bergsma, E., Kroesbergen, E., & Van Luit, J. H. (2015).
executive functions and their contributions to complex “frontal Verbal and visual-spatial working memory and mathematical abil-
lobe” tasks: A latent variable analysis. Cognitive Psychology, ity in different domains throughout primary school. Memory &
41, 49–100. Cognition, 43, 367–368.
Möhring, W., Newcombe, N. S., & Frick, A. (2015). The relation Van der Elst, W., Hurks, P., Wassenberg, R., Meijs, C., & Jolles, J.
between spatial thinking and proportional reasoning in preschool- (2011). Animal verbal fluency and design fluency in school-aged
ers. Journal of Experimental Child Psychology, 132, 213–220. children: Effects of age, sex, and mean level of parental educa-
https​://doi.org/10.1016/j.jecp.2015.01.005. tion, and regression-based normative data. Journal of Clinical
Neuenschwander, R., Röthlisberger, M., Cimeli, P., & Roebers, C. and Experimental Neuropsychology, 33(9), 1005–1015. https​://
M. (2012). How do different aspects of self-regulation pre- doi.org/10.1080/13803​395.2011.58950​9.
dict successful adaptation to school? Journal of Experimental Verdine, B. N., Irwin, C. M., Golinkoff, R. M., & Hirsh-Pasek, K.
Child Psychology, 113(3), 353–371. https​://doi.org/10.1016/j. (2014). Contributions of executive function and spatial skills
jecp.2012.07.004. to preschool mathematics achievement. Journal of Experimen-
Newcombe, N. S., & Shipley, T. F. (2015). Thinking about spatial tal Child Psychology, 126, 37–51. https​://doi.org/10.1016/j.
thinking: New typology, new assessments. In Studying Visual and jecp.2014.02.012.
Spatial Reasoning for Design Creativity (p. 179–192). Dordrecht: Wechsler, D. (2003). Wechsler Intelligence Scale for Children (4th ed.).
Springer. https​://doi.org/10.1007/978-94-017-9297-4_10 San Antonio: The Psychological Corporation.
O’brien, R. M. (2007). A caution regarding rules of thumb for variance Wolfgang, C. H., Stannard, L. L., & Jones, I. (2001). Block play
inflation factors. Quality & Quantity, 41(5), 673–690. performance among preschoolers as a predictor of later school
Peng, P., Wang, T., Wang, C., & Lin, X. (2018). A meta-analysis on achievement in mathematics. Journal of Research in Childhood
the relation between fluid intelligence and reading/mathematics: Education, 15(2), 173–180. https​://doi.org/10.1080/02568​54010​
Effects of tasks, age, and social economics status. Psychological 95949​58.
Bulletin, 145, 189–236. Wynn, K. (1992). Addition and subtraction by human infants. Nature,
Röthlisberger, M., Neuenschwander, R., Cimeli, P., & Roebers, C. M. 358(6389), 749–750. https​://doi.org/10.1038/35874​9a0.
(2013). Executive functions in 5-to 8-year olds: Developmental Yeniad, N., Malda, M., Mesman, J., van IJzendoorn, M. H., & Pieper,
changes and relationship to academic achievement. Journal of S. (2013). Shifting ability predicts math and reading performance
Educational and Developmental Psychology, 3, 153–167. https​ in children: A meta-analytical study. Learning and Individual Dif-
://doi.org/10.5539/jedp.v3n2p​153. ferences, 23, 1–9. https​://doi.org/10.1016/j.lindi​f.2012.10.004.
Röthlisberger, M., Neuenschwander, R., Michel, E., & Roebers, M.
(2010). Exekutive Funktionen: Zugrundeliegende kognitive Publisher’s Note Springer Nature remains neutral with regard to
Prozesse und deren Korrelate bei Kindern im späten Vorschu- jurisdictional claims in published maps and institutional affiliations.
lalter. Zeitschrift für Entwicklungspsychologie und Pädagogische
Psychologie, 42(2), 99–110.
Schmitt, N. (1996). Uses and abuses of coefficient alpha. Psy-
chological Assessment, 8(4), 350–353. https ​ : //doi.
org/10.1037/1040-3590.8.4.350.

13
Psychological Research is a copyright of Springer, 2021. All Rights Reserved.