s10212-024-00866-x

European Journal of Psychology of Education (2024) 39:2969–2989
https://doi.org/10.1007/s10212-024-00866-x
German and Taiwanese secondary students’ mathematical

modelling task value profiles and their relation
to mathematical knowledge and modelling performance
Kai‑Lin Yang1 · Janina Krawitz2 · Stanislaw Schukajlow3 · Chai‑Ching Yang1 ·

Yu‑Ping Chang4
Received: 26 August 2023 / Revised: 23 May 2024 / Accepted: 7 June 2024 /

Published online: 21 June 2024
© Instituto Universitário de Ciências Psicológicas, Sociais e da Vida 2024
Abstract
Based on expectancy-value theory, this study adopted a person-centred approach to explore
the heterogeneous profiles of secondary German and Taiwanese students’ mathemati-
cal modelling task values, and examined the differences in their mathematical modelling
performance, controlling for the variable of intra-mathematical knowledge among the het-
erogeneous profiles. Authors conducted a survey study of 452 ninth graders (201 German
students and 251 Taiwanese students). The results showed that German and Taiwanese stu-
dents respectively displayed three profiles of mathematical modelling task values: a) mod-
erate utility and moderate interest/attainment, b) high utility but low interest/attainment,
and c) low utility but high interest/attainment. Furthermore, different profiles of mathe-
matical modelling task values showed significant differences in mathematical modelling
performance for Taiwanese students but not for German students, even after removing the
variable of intra-mathematical knowledge. This study advances the understanding of stu-
dents’ mathematical modelling task values and its relation with their mathematical model-
ling performance by the expectancy-value model of achievement motivation and person-
centred analyses, and sheds light on the learning and teaching of mathematical modelling.
Keywords Mathematical modelling · Task value · Latent profile analysis · Motivation
* Yu‑Ping Chang
ypchang@mail.ntue.edu.tw
1
National Taiwan Normal University, Taipei, Taiwan
2
University of Cologne, Cologne, Germany
3
University of Münster, Münster, Germany
4
Department of Education, National Taipei University of Education, No. 134, Sec. 2, Heping East
Rd, 10671 Taipei, Taiwan
13
Vol.:(0123456789)
2970 K.-L. Yang et al.
Introduction
Mathematical modelling has been viewed as promising education reform for solving math-
ematical modelling problems which require students to associate and transfer between the
real and mathematical worlds. This process is conducive to students becoming effective
real-world problem solvers in society in the future (Blum & Leiß, 2007; Gravemeijer et al.,
2017). In addition to studying how students effectively solve mathematical modelling tasks
from a cognitive perspective, the inclusion of the motivational perspective would make the
research of learning and teaching mathematical modelling more comprehensive (Cevikbas
et al., 2022). When investigating students’ emotion and motivation in solving mathemati-
cal modelling problems, prior studies focused less on subjective task values than on self-
efficacy, attitudes and beliefs (Chamberlin & Sriraman, 2019).
Subjective task value has been recognized as one factor influencing engagement and
performance according to the situated expectancy-value theory (SEVT) (Eccles & Wig-
field, 2002). Studies (Rellensmann & Schukajlow, 2017; Krawitz & Schukajlow, 2018)
have investigated students’ interest in problems without and with a connection to reality,
which is closely related to one component of task value (Eccles et al., 1983). Contrary
to general expectations, German students’ interest in modelling problems was lower than
their interest in intra-mathematical problems. To increase student interest in mathematical
modelling, Krawitz et al. (2022) further found that reading comprehension prompts can
significantly increase interest in mathematical modelling tasks for German students but not
for Taiwanese students. The result also reminds us that individuals’ motivations, perfor-
mance and relationships are influenced by socio-cultural contexts and factors.
In recent decades, there have been many studies on training strategies for mathematical
modelling courses in many countries (Cevikbas et al., 2022), especially in Germany, but
there have been few in Taiwan. Another reason for the comparison of German and Tai-
wanese students in this study is the differences between curricula in these two countries. In
Germany, modelling competence is adopted as one of the six competencies in the national
mathematics curriculum (KMK, 2004), while it is not the case in Taiwan’s current math-
ematics curriculum (Ministry of Education in Taiwan, 2018). Nonetheless, it is emphasized
to increase the connection of the mathematical problem solving in real-life situations (Min-
istry of Education in Taiwan, 2018). That is, mathematical modelling is still an issue of
concern for developing student mathematical competence in Taiwanese school education.
Thus, based on the SEVT in achievement-related choices and performances (Eccles &
Wigfield, 2002), authors considered a socio-cultural perspective on such diversity, and the
differences between German and Taiwanese curricula which could induce different profiles
of mathematical modelling task values (MMTVs) in the two countries, and then impact
students’ mathematical modelling performance. Studies on students’ mathematical task
values and MMTVs have mostly adopted variable-centred approaches to examine the cor-
relations among factors or the relationships between predictors and consequences for the
overall sample. However, these approaches ignore “the integration of motivational qualities
to shape academic outcomes and to consider simultaneous variations among multiple moti-
vational indicators within students” (Fong et al., 2021, p. 1).
Students with different backgrounds and experiences may not follow similar patterns in
the components of task values. Furthermore, the results generated by the variable-centred
approaches may not provide sufficient information for researchers and teachers to custom-
ize interventions for different student groups. The objectives of this study were to exam-
ine whether MMTVs were combined into heterogeneous profiles, and whether students of
13
German and Taiwanese secondary students’ mathematical modelling… 2971
some profiles were better than students of others in solving mathematical modelling tasks.
Specifically, the purpose of this study was twofold: (1) to explore the number and char-
acteristics of the heterogeneous profiles of German and Taiwanese secondary students’
MMTVs, and (2) to examine the differences between MMTV profiles in intra-mathemat-
ical knowledge, in mathematical modelling performance and in mathematical modelling
performance after controlling for intra-mathematical knowledge.
Conceptual background
Mathematical modelling and intra‑mathematical knowledge
Mathematical modelling is a cyclical process moving from the real world to the mathemat-
ical world and back to the real world, passing through understanding the task to construct
a situational model, simplifying and structuring the situational model as the real-world
model, mathematizing the real-world model to find a mathematical result, interpreting the
mathematical result, and validating, refining as well as presenting the result and the model-
ling process (Blum & Leiß, 2007; Galbraith & Stillman, 2006; Schukajlow et al., 2012).
To solve mathematical modelling tasks requires various abilities, for example, the ability
to make assumptions when dealing with real-world tasks, the ability to transfer real-world
tasks to the mathematical world, and the ability to apply intra-mathematical knowledge
when solving problems (see Cevikbas et al., 2022; Chang et al., 2020).
Intra-mathematical knowledge is a necessary precondition for success in modelling
(Krawitz & Schukajlow, 2018; Niss & Blum, 2020). Students need sufficient intra-math-
ematical knowledge to set up an appropriate mathematical model, and to perform the
necessary calculations to come to a mathematical result. However, solving mathematical
modelling problems comprises abilities that differ from solving intra-mathematical prob-
lems. Chang et al. (2020) showed that German and Taiwanese ninth-graders have particu-
lar strengths and weaknesses. Although Taiwanese learners had higher intra-mathemati-
cal abilities, when learners from both countries had the same level of intra-mathematical
knowledge, the German students were found to have higher modelling performance, indi-
cating that modelling abilities differ from intra-mathematical abilities.
Take one modelling situation of this study, the Kite task, as an example (see Fig. 1). The
situation of the modelling problem is that two friends bought a kite and played with it in a
meadow near a power line. The relevant information regarding the length of the kite line,
the height of the power line, and in what circumstance the kite line is at its full length is
given as text (see more details in Chang et al., 2020). In order to solve the two modelling
problems for the situation, problem solvers need to construct a model to fit the conditions
given in the situation, and make some numerical or nonnumerical assumptions to apply
intra-mathematical knowledge, the Pythagorean theorem, to solve the problems.
Task values
One critical approach for investigating achievement-related motivation in mathematics

education is the application of the expectancy-value theory (Schukajlow et al., 2023;
Eccles & Wigfield, 2002). Individuals with stronger expectancy beliefs that they will do
well on the task are more likely to have better performances, while value beliefs can bet-
ter predict their engagement and task choice (Eccles & Wigfield, 2002). According to
13
Fig. 1 Mathematical modelling situation for Kites (retrieved from Chang et al., 2020)
expectancy-value theory, subjective task values are assumed to shape students’ motiva-
tion and then affect their behaviours (Eccles-Parsons et al., 1983).
Concerning the interaction between individuals’ beliefs about the involved tasks and
their task performances, Eccles et al. (1983) identified four main components of subjec-
tive task value: utility, cost, interest, and attainment. Utility value means the perceived
usefulness of engagement in and achievement of a task for reaching an individual’s
short- or long-term goals (Eccles, 2005). For instance, students have high utility value
of problem solving if they perceive it as being useful to get high scores in the mathemat-
ics course and to perform very well in the future national examination.
Cost refers to the anticipated effort one will need to put into completing a task, and
what the individual has to give up to do a task (e.g. Do I do my homework or play out-
side?) (Eccles, 2005). That is, cost describes personal perceptions of the negative con-
sequences ascribed to engaging in a given task. Cost has further been differentiated into
three types: effort cost, opportunity cost and psychological cost (Eccles et al., 1983).
Herein, authors focus on opportunity cost as modelling activities have not been pop-
ularly emphasized in mathematics textbooks and national examinations. For instance,
taking time to engage in modelling problems reduces the time for other academic works.
Then, students will have high cost value if they only learn the fastest way to solve mod-
elling problems.
13
Interest value is viewed as one’s interest in engaging in a task. It is similar to interest

which includes both context-specific and enduring aspects (Hidi, 2000; Wigfield & Cam-
bria, 2010), and is related with Csikszentmihalyi (1997) idea of flow which is an optimal
mode of task engagement and is stored as positive feeling-related valences (Krapp, 2002).
If students have high interest value in a task, they will persist in accomplishing it for a long
period of time (Wigfield et al., 2009).
Attainment value is defined as the personal importance ascribed to engaging in a given
task and is viewed as being the most germane to academic performance (Eccles, 2005).
That is, students who have high attainment value in a task will value the accomplishment
of that task for their sense of themselves.
The latter two components of task value, interest and attainment values, which focus on
learning and mastery are intrinsic-related, while the first two components of task value,
utility value and cost, which focus on grades, approval from others and losses are extrinsic-
related (Duncan & McKeachie, 2005). The latter two may lead to high levels of intrinsic
motivation, while utility value can be tied to extrinsic motivation (Deci & Ryan, 1985;
Ryan & Deci, 2020). In addition, it has been found that opportunity cost in mathematics
is relatively high among success-oriented students and low among avoidance- and mas-
tery-oriented students (Tuominen et al., 2020, p. 10). Thus, cost is assumed to attribute to
extrinsic-related motivation, as students may engage more in a task if they care less about
what is given up to do it. The four components of task value have been justified as both
general and specific aspects (Part et al., 2020) and may vary across different mathematical
tasks (Schukajlow et al., 2012).
Relationships between mathematical modelling, task value and intra‑mathematical

knowledge
Many empirical studies have justified the relation of task value to mathematical perfor-
mance in general, but not to the specific area of mathematical modelling, although there
are studies on the relation of psychological factors, for example motivation, self-efficacy,
attitudes and beliefs, to mathematical modelling performance in particular (Cevikbas et al.,
2022; Schukajlow et al., 2023). The need for these research results from the importance of
these problems, which are stronger related to reality but usually more difficult than (stand-
ard) word problems (Verschaffel et al., 2020). There are few studies that have explored the
relation of task value to mathematical modelling performance. Krawitz and Schukajlow
(2018) specifically investigated German secondary students’ task value in different kinds of
problems and found that the students valued modelling problems less than dressed-up word
problems and mathematical problems.
Task value is elastic and highly dependent on past and present learning experiences
(Eccles, 2009). That is why the expectancy-value model of achievement choice and perfor-
mance highlights that current performance can become one’s previous achievement-related
experience, which in turn influences one’s task value and further performance (Wigfield
& Eccles, 2020). According to the model, task value and previous and further achieve-
ment-related performances are positively related. Hence, students with high task value
can be expected to have high intra-mathematical knowledge, which is recognized as previ-
ous achievement-related experience required to solve mathematical modelling problems,
and then high modelling performace. However, this relationship might differ for different
groups of students categorized by their MMTVs. It is still an open question how MMTVs
are related to intra-mathematical knowledge and modelling performance.
13
Differences of achievement, motivation and their relationship between east

and west
Researchers of achievement and motivation have been arguing that the higher the self-
concept or belief in one’s abilities, the higher one’s academic achievement (Schunk &
Pajares, 2009). However, according to released results of international comparative studies
in educational achievements such as TIMSS and PISA, it has consistently been shown that
most East Asian students have higher academic performances in mathematics than Western
students, but they often do not report higher motivation in doing mathematics (He & van
de Vijver, 2016). Nevertheless, the correlation between interest and performance in math-
ematics is stronger in East Asian (e.g. Korea, Taiwan, Japan, Hong Kong) when compared
to Western countries (OECD, 2013; Wang et al., 2020).
Furthermore, Zhu and Leung (2011) analysed the relationship between East Asian and
Western students’ achievement and motivation from TIMSS 2003 data in mathematics.
Regarding the motivation scale, there were seven relevant items of students’ enjoyment and
value in the learning of mathematics and structured after factor and item analysis. They
found that generally the contribution of intrinsic-related motivation (the factor of pleasure-
oriented motivation) to mathematical achievement in East Asian (Japan, Korea, and Tai-
wan) educational systems was higher than in Western (Australia, England, the Netherlands,
and U.S.) educational systems. The extrinsic-related motivation (the factor of productiv-
ity-oriented motivation) positively contributed to mathematical achievement in East Asian
(Hong Kong, Japan, Korea, and Taiwan), while it was not the case for Western educational
systems. It implies that motivation might influence achievements differently in Eastern and
Western countries.
Besides the difference in the relationship between achievement and motivation in two
different cultures, Randel (2001) found that East Asian students were more likely to report
that mathematics was useful for their future careers, while Western students were more
likely to report that mathematics was useful for problem solving. It seemed that although
students from both cultures extrinsically valued mathematics, that is the utility of math-
ematics, they had different foci in their utility value. As utility value is one value compo-
nent, authors assume that Eastern and Western students might have different profiles of
task values.
Research questions and hypotheses
Previous studies mainly employed variable-centred approaches (e.g. structural equation

modelling) to investigate factors influencing and being influenced by students’ task val-
ues. Furthermore, prior research, based on expectancy-value theory, demonstrated that
cultural milieu and socialization affected individuals’ task value beliefs, and the effects of
task value on achievement-related factors differ for individuals (Eccles & Wigfield, 2002).
Hence, this study employed a person-centred approach, latent profile analysis, to character-
ize secondary students’ MMTVs, and then explored their relation with intra-mathematical
knowledge and mathematical modelling performance.
The present study was guided by the following research questions regarding both
German and Taiwanese students: (1) What are characteristics of different MMTV pro-
files? (2) What are the differences in intra-mathematical knowledge and in mathematical
13
modelling performance among different MMTV profiles? (3) What are the differences
in mathematical modelling performance after controlling for the variable of intra-math-
ematical knowledge among different MMTV profiles?
The released results of PISA 2012 showed that the mean score of intrinsic motivation
for German students was lower than that for Taiwanese students (OECD, 2013). Krawitz
et al. (2022) study found that German students’ interest in modelling was also lower
than that of Taiwanese students. Therefore, we expected that German students’ MMTV
profiles would be different from those of Taiwanese students, and that there would be
lower levels of MMTV in Germany than in Taiwan. Based on expectancy-value theory
and on previous studies about the effects of intra-mathematical knowledge and inter-
est value on mathematical modelling performance, there were another two expectations.
One was that students with higher levels of MMTV would perform better on intra-
mathematical knowledge and mathematical modelling than students with lower levels of
MMTV. The other is that there would be significant differences among different MMTV
profiles in mathematical modelling performance when controlling for the effect of intra-
mathematical knowledge.
Methods
Participants and procedure
The present sample involved 452 ninth graders including 201 German students from
nine classes in high-track schools (German Gymnasium) and 251 Taiwanese students
from 12 classes in a junior high school. The German students had learnt the Pythago-
rean theorem in grade 9 and were expected to have had some experience of solving
modelling problems, while the Taiwanese students had learnt the topic of Pythagorean
theorem in grade 8 but did not have experience of solving modelling problems.
Data of this study were collected in the authors’ previous study. The previous study
was designed in two conditions (one with the treatment of reading comprehension ques-
tions and the other without) in both countries in order to analyse the effect of read-
ing comprehension questions on students’ modelling performance (see Krawitz et al.,
2022). Authors argue that the data collected in the two conditions were suitable to be
combined to explore the profiles of students’ MMTVs (see section Preliminary Analy-
sis). All students were firstly required to complete four mathematical modelling situ-
ations provided with or without reading comprehension questions and to fill in a task
value questionnaire within 60 min. Then, they worked on an intra-mathematical test for
20 min. This test was used to collect indications of their intra-mathematical knowledge
of the Pythagorean theorem.
Measures
The instruments in this study included items on intra-mathematical knowledge of the

Pythagorean theorem, four modelling situations, such as the aforementioned description of
the Kite Problem in Fig. 1, and 16 task-specific value items. For each modelling situation,
there were two modelling problems and four task-specific value items.
13
Intra‑mathematical knowledge test
Students’ intra-mathematical knowledge was measured by 10 items including conceptual

knowledge, standard procedural knowledge, and comprehensive procedural knowledge
(Chang et al., 2020) on the topic of the Pythagorean Theorem. The intra-mathemati-
cal knowledge was coded for the correctness of solutions (0 = false or missing solution,
1 = accurate solution). The scale’s internal consistency (Cronbach’s alpha) was 0.73 for
German students and 0.77 for Taiwanese students.
Modelling test
Students’ modelling performance – assessed via constructing a real model – was measured
by the eight problems from four modelling situations (Krawitz et al., 2022). Two coders
were involved in scoring the German test booklets, and six coders coded the Taiwanese part
of the sample. At least 20% of the test booklets in each country were used to calculate inter-
coder reliability. Two coders scored the solutions for each item. The coding was carried out
by university students in Germany and graduate students in Taiwan. All of the coders com-
pleted a training session and received a coding manual to give their scores (Krawitz et al.,
2022) according to the developed coding scheme (Chang et al., 2020). The coding scheme
is composed of two dimensions: the adequacy of mathematical knowledge (no/unsuitable/
suitable) in solving the mathematical model and the completeness of information (incor-
rect/incomplete/complete) in forming the real model. The various scores (from 0 to 4) are
crossed by the scales of these two dimensions (see Chang et al., 2020). When differences
occurred, the coders discussed their judgments and made a consensual decision to choose
one code. For coding the German participants, the intercoder agreement between two cod-
ers (Cohen’s κ) was 0.69 or higher. For coding the Taiwanese participants, the intercoder
agreement between two coders (Cohen’s κ) was 0.80 or higher. The intercoder agreement
results indicated an acceptable level of agreement. The scale reliability (Cronbach’s alpha)
for measuring the construction of a real-world model was 0.62 for German students and
0.72 for Taiwanese students. The average score of the construction of a real-world model
was presented by the mean of eight problems as an indication of modelling performance.
Task value scales
The construct of task value was assessed in terms of the utility, cost, interest and attain-
ment components of MMTVs. For each of the four modelling situations, after working on
the two problems of each modelling situation, four task-specific value items were rated
using a 5-point Likert scale from 1—strongly disagree to 5—strongly agree. Taking the
Kite situation as an example, the task-specific value items included (1) If tasks such as the
Kite task are not included in the entrance (qualification) examination, I don’t want to solve
it (Utility value); (2) I want to know the solution to the Kite situation in the fastest way
only (Cost); (3) It was interesting to work on this Kite situation (Interest value); (4) I think
it is valuable to learn how to solve this Kite situation (Attainment value). Value scales were
built by aggregating the items across four modelling situations to an arithmetical mean.
Consequently, each value scale included four items which differed in the situation in which
students had to rate their value.
13
Data analysis
Preliminary analysis
Authors conducted a preliminary analysis of data and tested the differences in intra-math-
ematical knowledge between the two conditions (with and without reading comprehen-
sion). The results indicated that there were no differences between German students from
the two conditions (with reading comprehension questions: M = 0.70, SD = 0.30; without
reading comprehension questions: M = 0.68, SD = 0.32; t(199) = 0.40, p = 0.69) as well as
between Taiwanese students from the two conditions (with reading comprehension ques-
tions: M = 0.66, SD = 0.40; without reading comprehension questions: M = 0.72, SD = 0.37;
t(249) = -1.14, p = 0.26). These results indicated that the two designed conditions could be
combined for the profile analysis.
Measurement invariance for mathematical modelling task value
When comparing German and Taiwanese students’ MMTVs, establishing measure-

ment invariance across groups is “a logical prerequisite to conducting substantive cross-
group comparisons” (Vandenberg & Lance, 2000, p. 4). That is, authors have to justify
the assumption of metric measurement invariance via a multigroup confirmatory factor
analysis. Considering low missing data, full information maximum likelihood and robust
maximum likelihood estimator were used. Guidelines for determining model fit included
the comparative fit index (CFI) and Tucker-Lewis Index (TLI) exceeding 0.9 (Bentler &
Bonnet, 1980), and root mean square error of approximation (RMSEA) below 0.08 (Mac-
Callum et al., 1996).
The metric model in which item loadings were constrained to be equal across sam-
ples showed adequate fit (see Table 1; χ2 (212) = 494.34, CFI = 0.90, TLI = 0.88,
RMSEA = 0.08) and negligible change in fit statistics relative to the configural model
which imposed no equality constraints (CFI reduced by 0.005, TLI by 0.009, RMSEA by
0.001; see Chen, 2007; Little, 1997).
The above combined multigroup confirmatory factor analysis (CFA) in Table 1 shows
that the German and Taiwanese students’ responses to each mathematical modelling task
value item were comparable. Latent correlations among the four components are presented
in Table 2 and the descriptive statistics are in Table 3. The latent correlations are different
for the German and Taiwanese students, particularly the correlations among cost and the
other three components. Both skewness and kurtosis values with an absolute value less
than 1 are considered excellent for evaluating normality (George & Mallery, 2010). Table 3
shows that skewness and kurtosis values were acceptable for the normality assumption.
Data analysis for answering the research questions
To address the first research question, which was “What are the characteristics of different
MMTV profiles?”, Latent Profile Analysis (LPA) was conducted using the four compo-
nents of each participant to identify the best fitting number of profiles. LPA is a model-
based method for clustering individuals in unobserved groups. The goal is to probe into
how many latent classes can be identified in the data and to estimate their proportional
size and response profiles (Bauer, 2022). Within each profile, participants had similar
13
Table 1 Multigroup CFA for Modelling situations UL ME@ II@Germany / Taiwan

the metric invariance model: Germany /
Unstandardized item loadings Taiwan
(UL), measurement errors (ME)
and item intercepts (II) Item of Utility
If tasks such as ___ are not included in the entrance (qualifica-
tion) examination, I don’t want to solve it
Luggage and Posters 0.82 0.15/0.11 2.66/3.00
Fire Brigade 0.89 0.10/0.07 2.46/3.05
Kite 0.91 0.13/0.10 2.77/3.10
Parachuting 0.89 0.14/0.09 2.96/3.03
Item of Cost
I want to know the solution to the ___ task in the fastest way only
Fire Brigade 0.98 0.12/0.06 3.38/3.21
Kite 1.07 0.13/0.04 3.29/3.17
Parachuting 1.00 0.14/0.06 3.30/3.15
Item of Interest
It was interesting to work on this ___ task
Fire Brigade 0.83 0.13/0.09 2.84/2.92
Kite 0.89 0.13/0.04 2.69/2.82
Parachuting 0.87 0.10/0.07 2.48/3.02
Item of Attainment
I think it is valuable to learn how to solve this ___ task
Fire Brigade 0.82 0.12/0.06 3.37/3.19
Kite 0.94 0.09/0.05 3.26/3.07
Parachuting 0.89 0.10/0.07 3.06/3.15
All items were rated 1 = strongly disagree, to 5 = strongly agree
Table 2 Latent correlations Utility Cost Interest Attainment

among German and Taiwanese
samples (from multigroup
Utility - -0.10 -0.70*** -0.73***
CFA with metric invariance
constraints) Cost 0.19* - 0.28*** 0.34***
Interest -0.59*** 0.07 - 0.87***
Attainment -0.50*** 0.10 0.70*** -
German sample below diagonal; Taiwanese sample above diagonal

Total Ns = 452, *p < 0.05, **p < 0.01, ***p < 0.001
patterns of MMTVs. The criteria for selecting the best fitting number of profiles included
the Akaike Information Criterion (AIC; Akaike, 1974), the Bayesian Information Criterion
(BIC; Schwarz, 1978), the adjusted BIC (aBIC), the Vuong-Lo-Mendell-Rubin likelihood
ratio test (VLRT) and the Lo-Mendell-Rubin adjusted likelihood ratio test (aLRT) (Nylund
et al., 2007). Lower values of the AIC, BIC and aBIC, and higher entropy suggest a better
13
Table 3 Descriptive statistics for German and Taiwanese samples

German sample Taiwanese sample
M(SD) skewness kurtosis M(SD) skewness kurtosis
Utility 2.71 (1.04) 0.31 -0.59 3.05 (0.92) 0.07 0.28

Cost 3.35 (1.12) -0.16 -0.96 3.20 (0.98) -0.06 -0.18
Interest 2.58 (0.95) -0.02 -0.66 2.93 (0.88) -0.53 0.47
Attainment 3.19 (1.04) -0.24 -0.58 3.16 (0.86) -0.43 0.83
Listwise Ns = 201 German sample; 251 Taiwanese sample
fitting model. The significant VLRTs and aLRTs indicate that having one more profile
(k + 1 profiles) better fits data over the model without the additional profile (k profiles)
(Lo et al., 2001). If there were the same number of profiles for German and Taiwanese
students, a series of multigroup LPAs would be conducted for the overall sample, includ-
ing unconstrained, mean-constrained, mean- and variance-constrained, mean- and profile
size-constrained, and fully constrained models (Eid et al., 2003). The model with the low-
est AIC, BIC and Sample-Adjusted BIC (saBIC) values under the condition of significant
VLMR and aLRT were considered to be the best solution.
Once the best fitting model was identified, each student was assigned with the profile
membership estimated in that model. Regarding each country, after the best fitting solu-
tions were identified, a MANOVA was performed to examine the differences in the Lik-
ert scales of MMTVs between the profiles. Within each profile, one-way repeated meas-
ure ANOVA tests with Bonferroni post hoc comparisons were also adopted to examine
whether students’ MMTVs differed within each profile using the Likert scales.
To answer the second research question, which was “What are the differences in intra-
mathematical knowledge and in mathematical modelling performance among differ-
ent MMTV profiles?”, authors used an ANOVA test to examine the differences in intra-
mathemtical knowledge and in mathematical modelling performance. To address the third
research question, an ANCOVA analysis was conducted to compare mathematical model-
ling performance, as the dependent variable, among profiles, as an independent variable,
after controlling for students’ intra-mathematical knowledge as the covariate. All analyses
were performed using Mplus 8.8 and SPSS 25. The significance level was set to 5%.
Results
Profiles of secondary students’ mathematical modelling task value
Table 4 presents the results of a series of LPA with two- to four-group solutions for Ger-
man and Taiwanese participants. For each country, the three-class and four-class model had
similar values of AIC, BIC, aBIC and entropy, but the VLMR and aLRT for the four-class
model was not significant (Table 4). The favourable p-values for the VLMR and aLRT
indicated that a four-class model did not fit significantly better than a three-class model.
Accordingly, the three-class model was selected for each country, although the four-group
solution had the lowest AIC, BIC, and aBIC values among the two- to four-group models.
13
Table 4 Results of a series of latent profile analyses

Germany Taiwan
No. of group 2 3 4 2 3 4
AIC 2244.09 2217.61 2214.29 2430.77 2300.45 2254.29

BIC 2287.03 2277.07 2290.27 2476.6 2363.91 2335.38
aBIC 2245.84 2220.04 2217.4 2435.38 2306.85 2262.46
Entropy 0.72 0.79 0.8 0.94 0.89 0.89
VLRT -1168.32*** -1109.04* -1090.81 -1322.84*** -1202.38* -1132.23
aLRT 114.24*** 35.15* 12.83 232.50*** 135.41* 54.2
Nth group
1st 68(33.8%) 107(53.2%) 58(28.9%) 37(14.7%) 170(67.7%) 158(63.0%)
2nd 133(66.2%) 33(16.4%) 28(13.9%) 214(85.3%) 36(14.3%) 25(10.0%)
3rd 61(30.4%) 107(53.2%) 45(17.9%) 54(21.5%)
4th 8(4.0%) 14(5.6%)
*p < 0.05, **p < 0.01, ***p < 0.001
Table 5 Model fit indices of multigroup LPA models for the overall sample
Model Unconstrained Mean-constrained Mean- and Mean- and profile Fully constrained
variance-con- size-constrained
strained
AIC 4990.88 5056.95 5063.93 5053.16 5153.14

BIC 5208.90 5225.61 5183.23 5213.59 5264.21
saBIC 5040.70 5095.49 5091.20 5089.82 5178.52
As shown in Table 5, the unconstrained model had the lowest AIC, BIC and saBIC val-
ues. For the unconstrained, partial constrained, and fully constrained models, within profile
mean, variance and profile size, were allowed to vary across samples. This indicated that
means, variances or sizes of the derived profiles differed between the German and Taiwan-
ese students’ MMTV. It was therefore justified to investigate MMTV profiles separately for
each country.
Table 6 reports the means of the four scales of MMTVs, and Fig. 2 shows students’
MMTVs of the three profiles in Germany and Taiwan.
Significant differences in students’ MMTVs were found among the three profiles for
German students (F (8, 392) = 39.690 (Pillai’s trace), p < 0.001), and for Taiwanese stu-
dents (F (8, 492) = 48.616 (Pillai’s trace), p < 0.001). Because the Box test highlighted that
the data lacked homogeneity for German students (F (20, 38,470.518) = 2.577, p < 0.001)
and for Taiwanese students (F (20, 38,283.046) = 3.589, p < 0.001), we used the Games-
Howell post hoc test for multiple comparison. For both German and Taiwanese students,
the tests revealed that the difference in the utility, attainment and interest components
of MMTVs between any two profiles was statistically significant. The result supported
that for both countries, Profile 2 had the highest mean score for the utility component of
MMTVs, while Profile 3 had the lowest mean score. For intrinsic-related MMTVs (inter-
est and attainment), Profile 2 had the lowest mean score, Profile 1 had a moderate level,
and Profile 3 had the highest mean score. Profile 2 demonstrated high utility value (U)
13
Table 6 Statistical measurements of task values for each profile in Germany and Taiwan
Germany Taiwan
Profile** 1 (n = 107, 2 (n = 33, 3 (n = 61, 1 (n = 170, 2 (n = 36, 3 (n = 45,

53.2%) 16.4%) 30.4%) 67.7%) 14.3%) 17.9%)
Measurement M SE M SE M SE M SE M SE M SE
Utility (U) 2.84 0.09 3.61 0.19 2.00 0.09 3.10 0.05 4.09 0.17 2.02 0.10
Cost 3.53 0.10 2.97 0.22 3.22 0.15 3.34 0.06 2.46 0.21 3.29 0.18
Interest 2.41 0.04 1.10 0.03 3.66 0.06 2.99 0.03 1.33 0.07 3.97 0.08
Attainment 3.18 0.08 1.85 0.12 3.94 0.10 3.17 0.03 1.78 0.12 4.26 0.07
Knowledge 0.67 0.03 0.63 0.06 0.75 0.04 0.66 0.03 0.59 0.06 0.89 0.05
Performance 1.14 0.09 1.13 0.16 1.36 0.13 0.65 0.06 0.43 0.13 1.28 0.15
Utility focused on examination; Cost focused on the fastest way only; Interest focused on interesting;
Attainment focused on valuable
**Profile 1 demonstrates moderate U and intrinsic-related (Intrinsic) MMTVs including interest and attain-
ment; Profile 2 demonstrates high U/low Intrinsic; Profile 3 demonstrates low U/high Intrinsic
Fig. 2 Students’ task values of the three profiles in Germany and Taiwan
but low intrinsic-related MMTVs (Intrinsic), labelled as high U/low Intrinsic. Similarly,
Profile 3 was labelled as low U/high Intrinsic, and Profile 1 was labelled as mod U/mod
Intrinsic. For the cost component, German students in three profiles had no significant dif-
ference. Different from German students, Taiwanese students with low U/high Intrinsic and
students with mod U/mod Intrinsic had no significant differences in cost value. Both pro-
files had a significantly higher mean for cost value than Taiwanese students with high U/
low Intrinsic. It indicated that Taiwanese students with high utility value would be less in
agreement with learning in the fastest way only.
Regarding German students, the mod U/mod Intrinsic, high U/low Intrinsic, and low U/
high Intrinsic profiles, respectively, consisted of 107 (53.2%), 33 (16.4%), and 61 (30.4%)
ninth-grade students. Regarding Taiwanese students, the mod U/mod Intrinsic, high U/
low Intrinsic, and low U/high Intrinsic profiles, respectively, consisted of 170 (67.7%), 36
(14.3%), and 45 (17.9%) ninth-grade students. The Chi-square test of homogeneity showed
that the distributions were significantly different for the German and Taiwanese students
(χ2 (2) = 11.48, p = 0.003). There is a higher percentage of the low U/high Intrinsic profile
and a lower percentage of the mod U/mode Intrinsic profile for German students than for
Taiwanese students.
13
Differences in knowledge and modelling performance among MMTV profiles
Regarding German students, there were no significant differences in intra-mathematical

knowledge or in mathematical modelling performance among the three profiles (F (2,
198) = 0.737, p = 0.480; F (2,198) = 0.288, p = 0.750). Regarding Taiwanese students,
there were significant differences in intra-mathematical knowledge and in mathemati-
cal modelling performance among the three profiles (F (2, 248) = 8.064, p < 0.001; F (2,
248) = 12.067, p < 0.001). Bonferroni post hoc comparisons indicated that the low U/high
Intrinsic profile performed significantly better than the mod U/mod Intrinsic profile and the
high U/low Intrinsic profile on intra-mathematical knowledge as well as on mathematical
modelling for Taiwanese students. There were no significant differences in intra-mathemat-
ical knowledge or in mathematical modelling performance between the mod U/mod Intrin-
sic profile and the high U/low Intrinsic profile.
Differences in modelling performance controlling the variable of knowledge

among MMTV Profiles
Regarding both German and Taiwanese students, the assumptions of homogeneity of var-
iances and regression slopes for intra-mathematical knowledge among the profiles were
not violated (FDE = (2, 198) = 0.275, p = 0.760; FDE = (2, 195) = 0.870, p = 0.421; FTW = (2,
248) = 2.149, p = 0.119; FTW = (2, 245) = 1.611, p = 0.202). When controlling for the effect
of the intra-mathematical knowledge, there was no significant relation of MMTV profiles
to mathematical modelling performance for German students (FDE = (2, 197) = 0.265,
p = 0.767), but there was a significant relation for Taiwanese students (FTW = (2,
247) = 4.580, p < 0.01, η2partial = 0.036). Bonferroni post hoc comparisons indicated that the
low U/high Intrinsic profile performed significantly better than the mod U/mod Intrinsic
profile and the high U/low Intrinsic profile on mathematical modelling for Taiwanese stu-
dents. This means that German students’ MMTV profiles do not differ in their modelling
performance when controlling for the effect of intra-mathematical knowledge. The low U/
high Intrinsic profile of Taiwanese students performed the best among the three profiles of
Taiwanese students on mathematical modelling.
Discussion
Studies have revealed that four components of task values, namely utility, interest, attain-
ment value, and cost, predict students’ achievement, and suggested further investigation
of the possible differences between these components and their relation to achievement-
related outcomes (e.g., Eccles & Wigfield, 2002). This study followed this call and applied
a person-centred approach to classify and characterize students’ MMTVs based on the four
components of values, and analysed the relations among value components and achieve-
ment-related outcomes. Latent profile analysis was applied to explore the profiles of stu-
dents’ task value to overcome the insufficiency of variable-centred approaches applied in
the past. Furthermore, authors analysed the differences between profiles in intra-mathemat-
ical knowledge and mathematical modelling performance in one Eastern country (Taiwan)
and one Western country (Germany). The findings shed light on students’ motivational
13
strengths and weaknesses and can be used for customizing interventions to support stu-
dents’ motivation for the learning of mathematical modelling. Moreover, the findings
unveil the complexity of intra-country comparison in MMTVs.
One important result of this study is that German and Taiwanese secondary students
displayed three similar profiles of MMTVs: the moderate utility and moderate interest/
attainment profile, the high utility but low interest/attainment profile, and the low utility
but high interest/attainment profile. This finding does not support authors’ expectation that
German and Taiwanese students would differ in their MMTV profiles, and from previous
findings that revealed lower MMTVs for Taiwanese than for German students (Krawitz
et al., 2022; OECD, 2013). One explanation of this result might be that the differences
between countries are reflected not in the differences in the existence of different profiles,
but in the distribution of students across profiles. Indeed, the percentage of German stu-
dents in the low utility/high intrinsic value profile (about 30%) is higher than the percent-
age of Taiwanese students (about 18%), but the percentage of German students in the mod
utility/mod intrinsic value profile (about 53%) is lower than the percentage of Taiwanese
students (about 68%). This finding implies that motivational profiles across different cul-
tures may vary, and this should be examined in future studies.
For German students, the three profiles displayed similar cost, which was evaluated in
terms of students’ preference for learning the fastest way only. This may imply that cost,
compared to the other three components of task value, is less influential to classify German
students’ MMTVs. However, Taiwanese students with low interest and attainment but high
utility, reported lower cost compared to the other two profiles. This indicates that learners
from this profile may not be willing to engage in learning, even if it is the fastest way, if
the tasks are not directly relevant to examinations. One explanation of this finding is the
difference in educational systems in Germany and Taiwan. In Taiwan, ninth graders need
to take a comprehensive assessment which determines which high or vocational schools
they will enter. This high-stakes examination may be related students’ value and selection
of learning activities. This is an important finding, as it indicates the significance of the
social-cultural beliefs and behaviour suggested in Eccles’ expectancy-value model (Eccles
& Wigfield, 2002). To reduce Taiwanese students’ (especially the high U/low I profile)
negative appraisal of cost required to complete modelling problems, it was suggested to
include modelling problems in the national curriculum and high-stakes examinations. This
suggestion resulted from Taiwanese eighth graders’ low interest in mathematics, which was
ranked at the bottom in PISA studies (Mullis et al., 2020), and because of that their inter-
est in modelling problems is very difficult raised. Furthermore, effective interventions for
increasing task value in mathematics suggest to ask students to reflect on the relevance of
mathematics for their lives and to offer students mathematical tasks related to their future
profession (Gaspard et al., 2021; Rach & Schukajlow, 2024). It is also worth exploring how
these interventions can be adjusted for and affect students with different MMTV profiles.
Most of the German and Taiwanese students belonged to the mod U/mod Intrinsic pro-
file. Specifically, students in this profile had lower interest value for mathematical modelling
than the other components of task value. The prior research indicated that utility, attainment,
and interest values can positively correlate to students’ mathematical achievement, especially
interest value (Schukajlow et al., 2023; Wang et al., 2020). Based on expectancy-value the-
ory, authors also hypothesized a positive relationship between MMTVs and intra-mathemati-
cal knowledge as well as mathematical modelling performance. Nonetheless, this hypothesis
was confirmed for Taiwanese but not for German students. Practically, it implies that increas-
ing students’ interest value for mathematical modelling might be an effective and efficient
way to develop most Taiwanese students’ mathematical modelling, but not that of German
13
students. However, this implication needs further research. Theoretically, the study provides
evidence to support the relationship between sociocultural factors and situated expectancy-
value (Eccles & Wigfield, 2002). With the practical consideration that modelling is impor-
tant for students’ future lives, and the fact that most students did not show high interest in
modelling, another suggestion is to include modelling problems in a national examination,
which is the concern of students in the high U/low Intrinsic profile, as this profile strongly
values tasks that are included in a national examination.
There is no doubt that modelling performance can be predicted by intra-mathematical
knowledge both from the theoretical and empirical perspectives (Chang et al., 2020; Blum &
Leiß, 2007). The novel important finding of this study is that, when controlling for the effect
of such a strong cognitive predictor, authors found that MMTVs were still significantly related
to students’ mathematical modelling performance in Taiwan, but not necessarily in Germany.
How can authors explain this result? Previous studies have verified the weaker relationships
between motivation/values and mathematical performance in Western than in Eastern coun-
tries using variable-centred approaches (e.g. Sansone & Harackiewicz, 2000; Wang et al.,
2020; Zhu & Leung, 2011). This study, using a person-centred approach, showed similar find-
ings and indicated the importance of the MMTV profiles for intra-mathematical knowledge
and mathematical modelling performance in Taiwan but not in Germany. These two countries
are viewed as representative examples of Eastern and Western countries.
The low U/high I profile of Taiwanese students had significantly higher modelling perfor-
mance when controlling for the effect of intra-mathematical knowledge than the other two
profiles. It means that both MMTV profiles and previous achievement-related experience
are important predictors of modelling performance for Taiwanese students. This echoes Zhu
and Leung’s (2011) finding of a positive relationship between Taiwanese students’ achieve-
ment and motivation; especially, intrinsic-related motivation is more predictive than extrinsic-
related motivation for mathematical achievement in Eastern countries. It also suggests that
MMTV profiles of Taiwanese students may need different motivational inventions and diffi-
culty levels of mathematical modelling activities by considering both their MMTVs and prior
intra-mathematical knowledge. Comparatively, the different MMTV profiles of German stu-
dents can work on modelling problems with similar difficulty levels of mathematical model-
ling activities as they displayed a similar level of intra-mathematical knowledge and math-
ematical modelling performance among the different MMTV profiles. However, again, further
research is needed to justify the findings. To instigate the motivation of students in different
profiles, modelling problems and teaching instruction can be different regarding emphasizing
the relevance of the context of modelling problems or the relation to the current students’ lives
(Gaspard et al., 2021).
Additionally, the scale reliability (Cronbach’s alpha) in the modelling test was lower among
German students than among Taiwanese students, which might be due to cultural differences
in students’ attitudes towards examinations. Taiwanese students may have the habit of taking
paper-and-pencil tests more seriously, even though it is just an informal quiz, while it may not
be the case for German students.
Conclusion and limitations
In the present study, three different MMTV profiles were identified for both German and
Taiwanese secondary students, including the moderate utility and moderate interest/attain-
ment profile, the high utility but low interest/attainment profile, and the low utility but
13
high interest/attainment profile. The three profiles of German students displayed similar
intra-mathematical knowledge and mathematical modelling performance, while the three
profiles of Taiwanese students displayed dissimilarly. The low U/high Intrinsic profile of
Taiwanese students had higher intra-mathematical knowledge as well as mathematical
modelling performance than the other two profiles.
This study has some limitations. Considering the problem types and mathematics con-
tent (Krawitz & Schukajlow, 2018), the results of this study cannot be generalized to all
modelling tasks, as this study included modelling problems that mathematically required
calculation of one side of a right-angled triangle that could be solved by using the Pythag-
orean theorem. Moreover, considering the limited samples, this study combined samples
of two treatments (with/without reading comprehension) in conducting the analysis. The
relationship between reading comprehension and MMTV can be further verified in future
studies.
Taking a person-centred approach, this study advances the current understanding of stu-
dents’ MMTVs, and contributes to research on the learning and teaching of mathemati-
cal modelling. Consequently, this study has important theoretical implications by suggest-
ing the stability of motivational profiles across cultures and confirming the importance of
social-cultural factors for achievement-related outcomes. Important practical implications
can be drawn for instruction in the classroom, which should take into account motivational
profiles for example by designing tasks that address different motivational profiles.
Acknowledgements This study was conducted in the framework of the Taiwanese-German research pro-
gram on cultural-societal influences on mathematics education (TaiGer).
Funding The meetings of the German and Taiwanese partners were funded by Deutsche Forschungsgemein-
schaft (DFG) and the National Science and Technology Council (NSTC, No. 110–2511-H-003–008-MY3),
allocated to Aiso Heinze (IPN Kiel, Germany) and Kai-Lin Yang (NTNU Taipei, Taiwan), respectively.
Declarations
Conflict of interest The authors declare no competing interests.
References
Akaike, H. (1974). A new look at the statistical model identification. IEEE Transactions on Automatic Con-
trol, 19(6), 716–723.
Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of covariance
structures. Psychological Bulletin, 88(3), 588–606.
Blum, W., & Leiß, D. (2007). How do students and teachers deal with mathematical modeling problems?
The example of Sugerloaf. In C. Haines, P. Galbraith, W. Blum, & S. Khan (Eds.), Mathematical mod-
elling: Education, engineering and economics—ICTMA12 (pp. 222–231). Horwood.
Bauer, J. (2022). A primer to latent profile and latent class analysis. Methods for researching professional
learning and development: Challenges, applications and empirical illustrations (pp. 243–268).
Springer International Publishing.
Cevikbas, M., Kaiser, G., & Schukajlow, S. (2022). A systematic literature review of the current discussion
on mathematical modelling competencies: State-of-the-art development in conceptualizing, measur-
ing, and fostering.Educational Studies in Mathematics, 109, 205–236.
Chamberlin, S. A., & Sriraman, B. (2019). Affect in mathematical modeling. Springer.
Chang, Y.-P., Krawitz, J., Schukajlow, S., & Yang, K.-L. (2020). Comparing German and Taiwanese sec-
ondary school students’ knowledge in solving mathematical modelling tasks requiring their assump-
tions. ZDM-Mathematics Education, 52, 59–72.
13
Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural
Equation Modeling: A Multidisciplinary Journal, 14(3), 464–504.
Csikszentmihalyi, M. (1997). Finding flow: The psychology of engagement with everyday life. Basic Books.
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior. Plenum
Press.
Duncan, T. G., & McKeachie, W. J. (2005). The making of the motivated strategies for learning question-
naire. Educational Psychologist, 40(2), 117–128.
Eccles J. S. (2005). Subjective task value and the Eccles et al. Model of achievement-related choices. In A.
J. Elliot, C. S. Dweck (Eds.), Handbook of competence and motivation (pp. 105– 121). Guilford Press.
Eccles, J. S. (2009). Who am I and what am I going to do with my life? Personal and collective identities as
motivators of action. Educational Psychologist, 44, 78–89.
Eccles, J., Adler, T. F., Futterman, R., Goff, S. B., Kaczala, C. M., & Meece, J. L. (1983). Expectations,
values, and academic behaviors. In J. T. Spence (Ed.), Achievement and achievement motives: Psy-
chological and sociological approaches (pp. 75–146). Freeman.
Eccles, J. S., & Wigfield, A. (2002). Motivation beliefs, values, and goals. Annual Review of Psychology,
53, 109–132.
Eccles -Parsons, J. S., Adler, T. F., Futterman, R., Goff, S. B., Kaczala, C. M., & Meece, J. L., et al.
(1983). Expectancies, values, and academic behaviors. In J. T. Spence (Ed.), Achievement and
achievement motivation (pp. 75–146). Freeman.
Eid, M., Langeheine, R., & Diener, E. (2003). Comparing typological structures across cultures by mul-
tigroup latent class analysis: A primer. Journal of Cross-Cultural Psychology, 34(2), 195–210.
Fong, C. J., Kremer, K. P., Cox, C.H.-T., & Lawson, C. A. (2021). Expectancy-value profiles in math and
science: A person-centered approach to cross-domain motivation with academic and STEM-related
outcomes. Contemporary Educational Psychology, 65, 101962. https://doi.org/10.1016/j.cedpsych.
2021.101962
Galbraith, P. L., & Stillman, G. (2006). A framework for identifying student blockages during transitions
in the modelling process. ZDM–Mathematics Education, 38(2), 143–162.
Gaspard, H., Parrisius, C., Piesch, H., Kleinhansl, M., Wille, E., Nagengast, B., Trautwein, U., & Hul-
leman, C. S. (2021). The potential of relevance interventions for scaling up: A cluster-randomized
trial testing the effectiveness of a relevance intervention in math classrooms. Journal of Educa-
tional Psychology, 113(8), 1507–1528.
George, D., & Mallery, P. (2010). SPSS for Windows step by step : A simple guide and reference, 17.0
update. Allyn & Bacon.
Gravemeijer, K., Stephan, M., Julie, C., Lin, F. L., & Ohtani, M. (2017). What mathematics education
may prepare students for the society of the future? International Journal of Science and Mathemat-
ics Education, 15(1), 105–123.
He, J., & van de Vijver, F. J. (2016). The motivation-achievement paradox in international educational
achievement tests: Toward a better understanding. In R. B. King & A. B. I. Bernardo (Eds.), The
psychology of Asian learners (pp. 253–268). Springer.
Hidi, S. (2000). An interest researcher’s perspective: The effects of extrinsic and intrinsic factors on
motivation. In S. Sansone & J. M. Harackiewicz (Eds.), Intrinsic and extrinsic motivation: The
search for optimal motivation and performance (pp. 309–339). Academic.
KMK—Kultusministerkonferenz. (2004). Bildungsstandards im Fach Mathematik für den Primarstufen-
bereich (Jahrgangsstufe 4). Beschluss vom 15.10.2004.
Krapp, A. (2002). Structural and dynamic aspects of interest development: Theoretical considerations
from an ontogenetic perspective. Learning and Instruction, 12(4), 383–409.
Krawitz, J., & Schukajlow, S. (2018). Do students value modelling problems, and are they confident
they can solve such problems? Value and self-efficacy for modelling, word, and intra-mathematical
problems. ZDM–Mathematics Education, 50, 143–157.
Krawitz, J., Chang, Y.-P., Yang, K.-L., & Schukajlow, S. (2022). The role of reading comprehension in
mathematical modelling: Improving the construction of a real-world model and interest in Germany
and Taiwan.Educational Studies in Mathematics, 109, 337–359.
Little, T. D. (1997). Mean and covariance structures (MACS) analyses of cross-cultural data: Practical
and theoretical issues. Multivariate Behavioral Research, 32, 53–76.
Lo, Y., Mendell, N. R., & Rubin, D. B. (2001). Testing the number of components in a normal mixture.
Biometrika, 88(3), 767–778. https://doi.org/10.1093/biomet/88.3.767
MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and determination of
sample size for covariance structure modeling. Psychological Methods, 1(2), 130–149.
Ministry of Education in Taiwan. (2018). Curriculum guidelines for the 12-year basic education—ele-
mentary school, junior high school, and upper secondary school: The domain of Mathematics.
13
Mullis I. V., Martin M. O., Foy P., Kelly D. L., Fishbein B. (2020). TIMSS 2019 international results in
mathematics and science. Boston College, TIMSS & PIRLS International Study Center website.
https://timssandpirls.bc.edu/timss2019/international-results
Niss, M., & Blum, W. (2020). The learning and teaching of mathematical modelling. Routledge.
Nylund, K. L., Asparouhov, T., & Muthén, B. O. (2007). Deciding on the number of classes in latent
class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation
Modeling: A Multidisciplinary Journal, 14(4), 535–569. https://doi.org/10.1080/107055107015753
96
OECD. (2013). PISA 2012 results: Ready to learn—students’ engagement, drive and self-beliefs (Vol.
III). PISA, OECD Publishing.
Part, R., Perera, H., Marchand, G., & Bernacki, M. (2020). Revisiting the dimensionality of subjective
task value: Towards clarification of competing perspectives. Contemporary Educational Psychol-
ogy. https://doi.org/10.1016/j.cedpsych.202.101875
Rach, S., & Schukajlow, S. (2024). Affecting task values, costs, and effort in university mathematics
courses: The role of profession-related tasks on motivational and behavioral states. International
Journal of Science and Mathematics Education, 22, 1013–1035.
Randel, B. F. (2001). Achievement values: A comparison of Asian American and European American
adolescents. Doctoral dissertation, University of Michigan.
Rellensmann, J. & Schukajlow, S. (2017). Does students’ interest in a mathematical problem depend
on the problem’s connection to reality? An analysis of students’ interest and pre-service teachers’
judgments of students’ interest in problems with and without a connection to reality. ZDM–Math-
ematics Education, 49, 367–378.
Ryan, R. M., & Deci, E. L. (2020). Intrinsic and extrinsic motivation from a self-determination theory
perspective: Definitions, theory, practices, and future directions. Contemporary Educational Psy-
chology, 61, 101860.
Sansone, C., & Harackiewicz (Eds.). 2000. Intrinsic and extrinsic motivation: The search for optimal
motivation and performance. Academic Press.
Schukajlow, S., Leiss, D., Pekrun, R., Blum, W., Müller, M., & Messner, R. (2012). Teaching meth-
ods for modelling problems and students’ task-specific enjoyment, value, interest and self-efficacy
expectations. Educational Studies in Mathematics, 79(2), 215–237.
Schukajlow, S., Rakoczy, K., & Pekrun, R. (2023). Emotions and motivation in mathematics education:
Where we are today and where we need to go. ZDM–Mathematics Education, 55, 249–267.
Schunk, D. H., & Pajares, F. (2009). Self-efficacy theory. In K. Wentzel & A. Wigfield (Eds.), Handbook
of motivation at school (pp. 35–53). Routledge.
Schwarz, G. (1978). Estimating the dimension of a model. The Annals of Statistics, 6(2), 461–464.
http://www.jstor.org/stable/2958889
Tuominen, H., Juntunen, H., & Niemivirta, M. (2020). Striving for success but at what cost? Subject-spe-
cific achievement goal orientation profiles, perceived cost, and academic well-being. Frontiers in Psy-
chology, 11. https://doi.org/10.3389/fpsyg.2020.557445
Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of the measurement invariance lit-erature:
Suggestions, practices, and recommendations for organizational research. Organizational Research
Methods, 2, 4–69.
Verschaffel, L., Schukajlow, S., Star, J., & Van Dooren, W. (2020). Word problems in mathematics educa-
tion: A survey. ZDM - Mathematics Education, 52, 1–16. https://doi.org/10.1007/s11858-020-01130-4
Wang, M.-T., Guo, J., & Degol, J. L. (2020). The role of sociocultural factors in student achievement moti-
vation: A cross-cultural review. Adolscent Research Review, 5, 435–450.
Wigfield, A., & Cambria, J. (2010). Students’ achievement values, goal orientations, and interest: Defini-
tions, development, and relations to achievement outcomes. Developmental Review, 30(1), 1–35.
Wigfield, A., & Eccles, J. S. (2020). 35 years of research on students’ subjective task values and motiva-
tion: A look back and a look forward. In A. J. Elliot (Ed.), Advances in motivation science (Vol. 7, pp.
161–198). Academic Press.
Wigfield, A., Tonks, S. M., & Klauda, S. L. (2009). Expectancy-value theory. In K. Wentzel & A. Wigfield
(Eds.),Handbook of motivation in school (pp. 55–76). Mahwah, NJ: Erlbaum.
Zhu, Y., & Leung, F. K. S. (2011). Motivation and achievement: Is there an East Asian model? International
Journal of Science and Mathematics Education, 9, 1189–1212.
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
13
Kai-Lin Yang. National Taiwan Normal University, Taipei, Taiwan
Current themes of research:

Students’ cognitive and motivational (e.g., interest, knowledge, performance, and task value) factors in
solving mathematical modelling
Most relevant publications in the field of psychology of education:
tions. ZDMMathematics Education, 52, 59–72.
and Taiwan. Educational Studies in Mathematics, 109, 337–359.
Janina Krawitz. University of Cologne, Cologne, Germany

Krawitz, J., & Schukajlow, S. (2018). Do students value modelling problems, and are they confident they
can solve such problems? Value and self-efficacy for modelling, word, and intra-mathematical prob-
lems. ZDM–Mathematics Education, 50, 143–157.
Stanislaw Schukajlow. University of Münster, Münster, Germany
Cevikbas, M., Kaiser, G., & Schukajlow, S. (2022). A systematic literature review of the current discussion
on mathematical modelling competencies: State-of-the-art development in conceptualizing, measur-
ing, and fostering. Educational Studies in Mathematics, 109, 205–236.
ondary 10.1007/s10212-024-00866-x school students’ knowledge in solving mathematical modelling
tasks requiring their assumptions. ZDMMathematics Education, 52, 59–72.
Krawitz, J., & Schukajlow, S. (2018). Do students value modelling problems, and are they confident they
can solve such problems? Value and self-efficacy for modelling, word, and intra-mathematical prob-
lems. ZDM–Mathematics Education, 50, 143–157.
Rach, S., & Schukajlow, S. (2024). Affecting task values, costs, and effort in university mathematics courses:
The role of profession-related tasks on motivational and behavioral states. International Journal of Sci-
ence and Mathematics Education, 22, 1013–1035. https://doi.org/10.1007/s10763-023-10413-7.
Rellensmann, J. & Schukajlow, S. (2017). Does students’ interest in a mathematical problem depend on the
problem’s connection to reality? An analysis of students’ interest and pre-service teachers’ judgments
of students’ interest in problems with and without a connection to reality. ZDM–Mathematics Educa-
tion, 49, 367–378.https://doi.org/10.1007/s11858-016-0819-3.
13
Schukajlow, S., Leiss, D., Pekrun, R., Blum, W., Müller, M., & Messner, R. (2012). Teaching methods for
modelling problems and students’ task-specific enjoyment, value, interest and self-efficacy expecta-
tions. Educational Studies in Mathematics, 79(2), 215–237.
Schukajlow, S., Rakoczy, K., & Pekrun, R. (2023). Emotions and motivation in mathematics education:
Where we are today and where we need to go. ZDM–Mathematics Education, 55, 249–267.
Chai-Ching Yang. National Taiwan Normal University, Taipei, Taiwan

No (She is new in the field)
Yu-Ping Chang. Department of Education, National Taipei University of Education, No. 134, Sec. 2, Heping
East Rd. 10671 Taipei, Taiwan. ypchang@mail.ntue.edu.tw
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a
publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript
version of this article is solely governed by the terms of such publishing agreement and applicable law.
13

s10212-024-00866-x

Uploaded by

Copyright:

Available Formats

s10212-024-00866-x

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

s10212-024-00866-x

Uploaded by

Copyright:

Available Formats

European Journal of Psychology of Education (2024) 39:2969–2989

German and Taiwanese secondary students’ mathematical

Kai‑Lin Yang1 · Janina Krawitz2 · Stanislaw Schukajlow3 · Chai‑Ching Yang1 ·

Received: 26 August 2023 / Revised: 23 May 2024 / Accepted: 7 June 2024 /

Keywords Mathematical modelling · Task value · Latent profile analysis · Motivation

Mathematical modelling and intra‑mathematical knowledge

One critical approach for investigating achievement-related motivation in mathematics

Interest value is viewed as one’s interest in engaging in a task. It is similar to interest

Relationships between mathematical modelling, task value and intra‑mathematical

Differences of achievement, motivation and their relationship between east

Research questions and hypotheses

Previous studies mainly employed variable-centred approaches (e.g. structural equation

Participants and procedure

The instruments in this study included items on intra-mathematical knowledge of the

Intra‑mathematical knowledge test

Students’ intra-mathematical knowledge was measured by 10 items including conceptual

Task value scales

Measurement invariance for mathematical modelling task value

When comparing German and Taiwanese students’ MMTVs, establishing measure-

Data analysis for answering the research questions

Table 1 Multigroup CFA for Modelling situations UL ME@ II@Germany / Taiwan

All items were rated 1 = strongly disagree, to 5 = strongly agree

Table 2 Latent correlations Utility Cost Interest Attainment

German sample below diagonal; Taiwanese sample above diagonal

Table 3 Descriptive statistics for German and Taiwanese samples

Utility 2.71 (1.04) 0.31 -0.59 3.05 (0.92) 0.07 0.28

Listwise Ns = 201 German sample; 251 Taiwanese sample

Profiles of secondary students’ mathematical modelling task value

Table 4 Results of a series of latent profile analyses

AIC 2244.09 2217.61 2214.29 2430.77 2300.45 2254.29

*p < 0.05, **p < 0.01, ***p < 0.001

AIC 4990.88 5056.95 5063.93 5053.16 5153.14

Profile** 1 (n = 107, 2 (n = 33, 3 (n = 61, 1 (n = 170, 2 (n = 36, 3 (n = 45,

Differences in knowledge and modelling performance among MMTV profiles

Regarding German students, there were no significant differences in intra-mathematical

Differences in modelling performance controlling the variable of knowledge

Conclusion and limitations

Kai-Lin Yang. National Taiwan Normal University, Taipei, Taiwan

Current themes of research:

Janina Krawitz. University of Cologne, Cologne, Germany

Chai-Ching Yang. National Taiwan Normal University, Taipei, Taiwan

Current themes of research:

Most relevant publications in the field of psychology of education:

Current themes of research:

Most relevant publications in the field of psychology of education:

You might also like

p < 0.05, p < 0.01, p < 0.001