Math anxiety affects career choices during development

Abstract

Background

Links between math anxiety and the choice of a math-intensive career might change over development and differ by gender. The study included three research populations: primary school (N = 87, 48 females, mean age = 10.2), high school (N = 107, 61 females, mean age = 15.7), and university students (N = 100, 53 females, mean age = 27.4). Students completed a math anxiety questionnaire and reported their desired career choice.

Results

Findings suggest that math anxiety directly predicted the career choice math intensity for high school and university students, but not primary school students. Gender had a direct effect on younger students, as female students attending primary and high school preferred careers with a lower math intensity. The effect of gender on career choice math intensity for university students was not direct but mediated by math anxiety.

Conclusions

It is crucial to identify young students with math anxiety and provide appropriate math anxiety reduction programs to reduce the cumulative effect of math anxiety on academic achievement and career choice.

Introduction

Math knowledge is positively linked with prosperity in our industrial, technological society (Peri et al., 2016). Therefore, there is increasing interest in the pursuit of a math-based career, for example, in the fields of science, technology, engineering, and math (STEM) (Amato, 2021). State-of-the-art research is trying to determine which variables can predict STEM careers (Chan, 2022; National Science Foundation, 2019; Organization for Economic Co-operation and Development [OECD], 2022; Rozek et al., 2019). Concurrent with the assertion that motivational–emotional factors can be better predictors of the pursuit of STEM careers than math aptitude (Dekhtyar et al., 2018; Wang & Degol, 2013), we focused on math anxiety as a predictor of the math intensity of a chosen career.

Math anxiety essentially refers to a spectrum of negative emotional responses to math information or activities (Ashcraft, 2002; Cipora et al., 2022), involving fear or worry (Richardson & Suinn, 1972) related to math stimuli in both educational (Lau et al., 2022) and everyday situations (Guzmán et al., 2023; Skagerlund et al., 2018). Math anxiety is a multidimensional construct involving elevated physiological activity (Eidlin Levy & Rubinsten, 2021; Qu et al., 2020), negative emotional responses (Cohen et al., 2021), and negative attitudes to math (Gunderson et al., 2011). The manifestations of math anxiety can include adverse reactions to math (Ashcraft, 2002), avoidance behavior (Daker et al., 2021; Jenifer et al., 2022), and low math-related self-confidence (Ahmed et al., 2012; Morán-Soto & González-Peña, 2022). Accordingly, math anxiety may be related to an avoidance of math learning and a reluctance to choose STEM courses and careers (Ahmed, 2018; Megreya & Al-Emadi, 2023; Morán-Soto & González-Peña, 2022). Math anxiety is accompanied by low math achievement across cultures (Barroso et al., 2020; Lau et al., 2022; Zhang et al., 2019) and developmental phases (Hart & Ganley, 2019; Hill et al., 2016; Tomasetto et al., 2021). However, it is important to note that high math anxiety is not automatically accompanied by low math achievement because of certain resilience factors, such as high motivation (Wang et al., 2015) or inhibition of anxiety symptoms (Lyons & Beilock, 2012). We investigated whether math anxiety was associated with the math intensity of career choices at three educational phases: primary school, high school, and university.

Do links between math anxiety and career choice change over time?

Math anxiety seems to increase over the years of schooling. In one study, 11% of elementary and middle school students reported high math anxiety levels (Devine et al., 2018), but this number may double or even triple among older students (Hart & Ganley, 2019; OECD, 2013). We aimed to expand existing knowledge on the association of math anxiety and career choice across gender at different educational phases. We used the social cognitive career theory (Brown & Lent, 2019; Kohen & Nitzan, 2021; Reinhold et al., 2018) to specify three educational phases.

The theory defines the first phase as the wishing phase (primary to junior high school). At this point, students may have an interest in a specific career, with no actual commitment to or decision-making towards career pursuit. To the best of our knowledge, the direct relations between math anxiety and career interests in primary school have not been investigated. However, recent research has found social-emotional factors, such as stereotype endorsement, can influence primary school students' interest in STEM-based subjects regardless of their achievements (Blažev et al., 2017), and students at this educational phase already have occupational gender-biased beliefs (Moè, 2018; Teliousi et al., 2020). Furthermore, positive attitudes towards STEM learning were positively associated with computational thinking (Sun et al., 2021) as well as math achievement (Gulemetova et al., 2022) among primary school students. Researchers have found links between math anxiety and math achievement even in the very first years of primary school (Gunderson et al., 2018; Tomasetto et al., 2021), and math anxiety has been found to interfere with the acquisition of new numerical knowledge (Tomasetto et al., 2021). Hence, as career interests among primary school students are related to their past experiences (Howard et al., 2015), it is plausible to hypothesize that avoidance of math activities will be evident as early as primary school, and children with high math anxiety may already show an interest in non-math-intensive careers.

The second phase is the planning phase (high school). In this phase, students set specific career goals, such as taking advanced math courses. Students who pursue STEM careers may make that choice during middle and high school when they discover a growing interest in mathematics and sciences (Babarović, 2022; Maltese & Tai, 2010). At the same time, math anxiety seem to increase for this age group (OECD, 2013), as does the association between math anxiety and low math achievements (Barroso et al., 2020; Zhang et al., 2019). Relevant to the current research, math anxiety among middle school students has been associated with interest in careers (Eidlin Levy et al., 2021; Huang et al., 2019) and the later pursuit of non-math-intensive careers (Ahmed, 2018). It also seems that high math anxiety can lead to an avoidance of math-intensive activities (Else-Quest et al., 2010; Jenifer et al., 2022). This avoidance has consequences: later on, low math achievements lessen the likelihood of acceptance into STEM studies (Kohen & Nitzan, 2021).

The third phase is the acting phase (higher education). Students make more important career-related decisions, such as choosing a major field of study. Recent research indicates that math anxiety is related to a reluctance to enroll in math courses (Soysal et al., 2022). One study found that math anxiety was a better predictor of STEM course avoidance than math ability among college students (Daker et al., 2021). As discussed earlier, this pattern may be the result of consistent negative links between math anxiety and interest in math-intensive careers, starting in primary school.

We were interested in the possible changing relations between gender, math anxiety, and career choice math intensity throughout the educational journey and thus focused on the wishing (primary school), planning (high school), and acting (university) phases. We also asked if gender might play a role in these decisions in any of the three phases.

Does gender account for the links between math anxiety and career choice?

Female students usually report higher math anxiety (Devine et al., 2012; Hart & Ganley, 2019; Hill et al., 2016; Morán-Soto & González-Peña, 2022), even though gender differences in numerical performance have decreased in the last few decades (Hyde & Mertz, 2009; Lindberg et al., 2010). Starting from primary school, female students are also less likely to be interested in or to enroll in math and science courses than male students (Babarović, 2022; Chan, 2022; Grosch et al., 2022; Watt, 2016). Consequently, regardless of their math competence, females are underrepresented in STEM departments in higher education and in STEM careers (National Science Foundation, 2019; OECD, 2022). While genetic–biological differences may account for higher math anxiety in females (Júlio-Costa et al., 2019; Wang et al., 2014), social agents, such as parents and teachers, have gender-biased expectations of math competence (Beilock et al., 2010; Gunderson et al., 2011; Lau et al., 2022), and these influence the development of math anxiety.

A few studies have investigated the impact of math anxiety on career choice while controlling for gender (Ahmed, 2018; Daker et al., 2021; Eidlin Levy et al., 2021; Huang et al., 2019; John et al., 2020; Megreya & Al-Emadi, 2023; Morán-Soto & González-Peña, 2022). Some have found that students (both boys and girls) who report consistent or increasing negative experiences with math are more likely to avoid math courses (Ahmed, 2018; John et al., 2020), and this, in turn, may result in avoidance of math-intensive careers. Recently, math anxiety was found to predict university students’ enrollment in STEM courses over and above math ability for both genders (Daker et al., 2021). Others have found a direct effect of math anxiety on STEM career interests among middle and high school girls but not boys (Eidlin Levy et al., 2021; Huang et al., 2019; Megreya & Al-Emadi, 2023). Thus, it is not clear whether the links between math anxiety and career choice apply only to females. Gender is typically found to be a predictor of career path (Chan, 2022; National Science Foundation, 2019; OECD, 2022) and math anxiety, which is more common among females, may moderate or mediate the relations between gender and career choice.

The study

We aimed to measure whether the links between math anxiety and the math intensity of a chosen career would differ by gender across the three educational phases. Based on the social cognitive career theory (Brown & Lent, 2019; Kohen & Nitzan, 2021; Reinhold et al., 2018), we used a cross-sectional design with three research populations. Our study was set in Israel; thus, the populations reflect the Israeli educational system.

The first population comprised primary school students in the 4th and 5th grades, the wishing step of social cognitive career theory (Brown & Lent, 2019; Kohen & Nitzan, 2021; Reinhold et al., 2018). We hypothesized that the career choices of 4th and 5th grade students are already likely to relate to former experiences (Howard et al., 2015), including math anxiety (Gunderson et al., 2018; Tomasetto et al., 2021). In Israel, 10th grade students, our second research population, enroll in oriented studies, and these choices may indicate their future career plans (Krill et al., 2019). This population represents the pre-action planning phase of social cognitive career theory, a period when students set goals to follow certain career paths. The third research population was university students. At this phase, students have chosen a major field of study, another indicator of career choice and one much closer to the actual career trajectory. This population is in the action phase (Brown & Lent, 2019; Reinhold et al., 2018). The focus on these particular populations enabled us to track trends in the relations between math anxiety and career choices across educational phases, from the wishing phase to the action phase (Brown & Lent, 2019; Reinhold et al., 2018).

Note that the education system in Israel includes universities, academic colleges, and technical institutes; student applications are based on the specific fields of study offered in each, and their acceptance rests on former academic achievements (Ayalon & Yogev, 2005). Although our sample included students in all three higher education systems, in what follows, we simply refer to university students.

We had two research questions. First, we asked if there are gender difference in math anxiety and/or career choice math intensity. Second, we asked what are the contributions of gender and math anxiety to career choice math intensity at different developmental phases? We hypothesized that females would opt for careers with a lower math intensity (Chan, 2022; Grosch et al., 2022; National Science Foundation, 2019; OECD, 2022). We further hypothesized that math anxiety levels would be higher for females than males, especially in later educational phases (Devine et al., 2012; Hart & Ganley, 2019; Hill et al., 2016), and that this math anxiety might, in turn, discourage students from pursuing math-intensive careers (Eidlin Levy et al., 2021; Huang et al., 2019; Megreya & Al-Emadi, 2023; Morán-Soto & González-Peña, 2022). However, some studies have found links between math anxiety and career choices for both genders (Ahmed, 2018; Daker et al., 2021). Thus, in a cross-sectional design, we tested the plausibility of math anxiety as a mediator or a moderator of the association between gender and career choice to get a more precise and consistent description of the links between variables.

Materials and methods

Participants and procedure

A total of 427 Israeli students participated in the study. Of these, 294 belonged to one of the three research groups based on age: 87 primary school, 107 high school, and 100 university students. Additional 133 participants were part of a post-experimental group to determine the load of math knowledge required for different academic fields of study (similar to former methodology (Ganley et al., 2018); we defined this as career choice math intensity; see “Measures” section).

We performed a statistical power analysis to calculate the required sample size for both variance and regression analyses. For an independent T test with alpha = 0.05 and power = 0.80, the calculated sample size for medium effect size (d = 0.5) (Cohen, 1992) using G*Power software was approximately N = 51 for each group. For regression analysis with two predictors (gender and math anxiety), each moderately correlating with the outcome variable (r = 0.3) (Cohen, 1992), with alpha = 0.05 and power = 0.80, the calculated sample size using G*Power software was approximately N = 50. As we potentially had six groups (3 developmental phases X 2 genders), the estimated sample size was 300. Therefore, we aimed to recruit approximately 100 participants for each developmental phase.

Research groups

The study included participants from three cross-sectional samples, representing different developmental phases.

The first sample included 87 4th–5th grade students, 48 female (mean age = 10 years and 2 months, SD = 0.9) and 39 male (mean age = 10 years and 2 months, SD = 0.7) students, attending five different classes. They completed a math anxiety questionnaire and reported their desired career choice.

The second sample included 107 9th–10th grade students, 61 females (mean age = 15 years and 7 months, SD = 0.4) and 46 males (mean age = 15 years and 7 months, SD = 0.4), attending four different classes. These students completed a math anxiety questionnaire and reported their major field of oriented study, representing their desired career choice.

For both primary and high school students, after parental contest was obtained, students completed the experimental tasks in small groups at school during the school day.

The third group included 100 undergraduate university students, 53 females (mean age = 26.43, SD = 6.02) and 47 males (mean age 28.67, SD = 5.65), recruited via online advertisements sent by social networks. To ensure variety in areas of study and academic skills, some students attended a university, while others attended an academic college or technical institute, as these have different acceptance thresholds and specialize in different fields of study yet provide equivalent diplomas in the Israeli academic system. To create a diverse and stratified sample (Enarson et al., 2004; Iliyasu & Etikan, 2021), we published information on our study throughout the country and hence were able to recruit participants from different institutions and geographic regions. Note that student age is relatively high in Israeli higher education, as mandatory military service (lasting approximately 3 years) is usually completed prior to enrollment.

All participants had completed both a math matriculation exam and a psychometric entrance test as a prerequisite for acceptance into higher academic studies. For our study, we asked students to complete a math anxiety questionnaire and to report their study department, representing their career choice. The study procedure lasted about 15 min, and participants were paid 5USD for their participation. They completed the tasks individually.

The study was carried out in accordance with the recommendations of the Ethics Committee of the University and the Chief Scientist Office of the Education Ministry, with written informed consent from university students and parental consent for primary and high school students.

Measures

Math anxiety: Abbreviated Math Anxiety Scale

Participants answered a translated version of the Abbreviated Math Anxiety Scale (AMAS) (Hopko et al., 2014), a 9-item questionnaire. The AMAS is frequently used to measure math anxiety levels and has been validated in various cultural populations and age groups, including primary school, secondary school (Martín-Puga et al., 2020), and university students (Cho, 2022; Primi et al., 2014). The questionnaire is designed to reflect the degree of anxiety experienced in a variety of math-related tasks and situations using a 5-point Likert-type scale (1 = low anxiety; 5 = high anxiety). To obtain the total score, we summed the scores for all questions (score range: 9–45, internal consistency reliability α = 0.85). See Table 1 for descriptive statistics.

Table 1 Descriptive statistics

Career choice math intensity

Career choice math intensity—primary school

Primary school students reported their desired career choice by answering the question: ‘What do you want to be when you grow up?’ Then, we calculated how math-demanding each profession was using a 1–10 scale adapted from former research (Eidlin Levy et al., 2021). In the previous study, 9th grade participants (N = 89) were asked to classify the degree of math proficiency required for each profession appearing in a list (1 = very low; 10 = very high). An average score represented the math proficiency for each occupation. In the present study, we coded students’ answers based on this previously compiled list of occupational math intensity, by allocating the present response to the same or closest occupation on the list. For instance, the profession of dancer was not seen as having high math demands (mean = 1.41), while engineer was seen as having high demands (mean = 9.34).

Career choice math intensity—high school and university

High school students reported their major field of oriented study, and university students reported their study departments by answering the question: ‘What is your current oriented study (for high school students) or major study department (for university students)?’ As most fields of study were similar across these groups, we created a list with all reported fields of study, distributed it on social media (Facebook), and asked volunteer participants (100 females, mean age 40.33, SD = 11.97) to classify the degree of math intensity required for each field of study (from 1 = very low to 10 = very high). This sample has ecological validity, as it represents conceptions of the math intensity of different occupations by members (non-students) of the society into which students will integrate after graduating. Furthermore, the over-representation of females in the sample matches the fact that as in other countries, female Israeli students are more likely than male students to enroll in higher education (OECD, ).

As we expected, the degree of math intensity required for literature studies was low (mean = 2.27) and the degree for engineering was high (mean = 8.67). For the full list of fields of study and their math intensity score, see Additional file 1: Appendix S1. See Table 1 for descriptive statistics.

Statistical plan

Research question 1

Our first objective was to determine whether gender was related to students’ math anxiety levels and/or the math intensity of their career choice. We conducted independent T tests for each developmental phase separately. Note that we analyzed each developmental phase separately, but did not compare them, as the construct of career choice math intensity is likely to be incomparable between groups, given the restrictions of the structure of the Israeli education system (e.g., primary school students are not sorted into fields of study, but high school and university students are).

To construct models that were as similar as possible across groups, we standardized math anxiety and math intensity of career choice for each developmental phase before performing the main analyses. We transferred raw scores into T scores (Weller, 1984). This allowed us to create a mutual score scale across developmental phases and reduce group variance.

Research question 2

The second objective was to determine the associations and predictive links between gender, math anxiety, and career choice math intensity across developmental phases. Specifically, we asked whether math anxiety moderated or mediated the relations between gender and career choice math intensity (Hayes, 2013). To address this question, we conducted linear regressions with gender and math anxiety as predictors of career choice as a first step. Next, we added the interaction between gender and math anxiety to the equation to create a moderation model.

We conducted separate analyses to determine whether gender was a significant predictor of math anxiety. Concurrent with mediation analysis assumptions, we further explored whether math anxiety mediated the relations between gender and career choice math intensity for cases when gender was a significant predictor of math anxiety and math anxiety was a significant predictor of career choice math intensity.

Interpretation of results

Bayes factors were used for all analyses to strengthen the robustness of the results (Vandekerckhove et al., 2018), as Bayes factors express the ratio between the evidence in favor of the research hypothesis and the null hypothesis. Based on the classification recommended by Wagenmakers et. al. (2018), we interpreted results using the following strategy: a Bayes factor below 1 supports the null hypothesis; a Bayes factor between 1 and 3 is inconclusive; a Bayes factor above 3 supports the research hypothesis.

Results

Research question 1: are there gender difference in the experimental variables (math anxiety, career choice math intensity)?

We conducted T tests to assess possible gender differences in math anxiety and career choice math intensity across developmental phases. As described on Table 2, there were no significant differences in math anxiety among female and male students attending primary school or high school. However, female students reported higher math anxiety levels than males at university, and Bayes results supported this finding. As for career choices, female primary and high school students were interested in careers with lower math intensity, although Bayes results were inconclusive for high school students. Interestingly, although female university students had higher math anxiety levels than males, there were no significant gender differences in career choice math intensity.

Table 2 Gender differences in math anxiety and career choice math intensity by group

Research question 2: what are the contributions of gender and math anxiety to career choice math intensity at different developmental phases?

We conducted regression analyses to explore the association and predictive links between math anxiety and career choice math intensity by gender and over time. Specifically, we asked whether math anxiety mediated or moderated the links between gender and career choice math intensity at each developmental phase (primary school, high school, university).

Primary school students

Moderation analysis: linear regression with gender and math anxiety as predictors of career choices indicated that gender (β = 0.276, t = 2.60, P = 0.011, BF10 = 3.74) but not math anxiety (β = − 0.048, t = 0.45, P = 0.653, BF10 = 0.48) significantly predicted career choice math intensity [F_{(2, 84)} = 3.77, P = 0.027, r² = 0.082]. Moderation analysis revealed that gender remained a significant predictor of career choice (β = 0.276, t = 2.59, P = 0.011, BF10 = 2.86), but math anxiety was not significant in the equation (β = − 0.106, t = − 0.71, P = 0.481, BF10 = 0.37), nor was the interaction between gender and math anxiety (β = − 0.082, t = 0.55, P = 0.583, BF10 = 0.41). The entire model was marginally significant [F_{(3, 83)} = 2.59, P = 0.058, r² = 0.053]. Therefore, we rejected the moderation model.

Mediation analysis: gender did not significantly predict math anxiety (β = − 0.157, t = − 1.46, P = 0.147, BF10 = 0.57). Therefore, we did not conduct further mediation analysis.

To conclude, for primary school students (see Fig. 1) gender had a direct effect on career choice, such that male students chose careers with higher math intensity than females, without the moderation of math anxiety.

Fig. 1

Links between gender, math anxiety and math intensity of career choice for primary school students. Illustration of gender and math anxiety as predictors of career choice math intensity for primary school students. A Gender (β = 0.276, t = 2.60, P = 0.011, BF10 = 3.74), but not math anxiety (β = − 0.048, t = − 0.451, P = 0.653, BF10 = 0.48), significantly predicted career choice [F_{(2, 84)} = 3.77, P = 0.027, r² = 0.082]. B Gender did not significantly predict math anxiety (β = − 0.157, t = − 1.46, P = 0.147, BF10 = 0.57)

High school students

Moderation analysis: both gender (β = 0.187, t = 2.01, P = 0.047, BF10 = 2.35) and math anxiety (β = − 0.240, t = 2.57, P = 0.011, BF10 = 5.35) significantly predicted career choice math intensity [F_{(2, 104)} = 0.62, P = 0.005, r² = 0.10]. Moderation analysis [F_{(3, 103)} = 3.72, P = 0.014, r² = 0.10] revealed a significant effect of gender (β = 0.187, t = 1.99, P = 0.048, BF10 = 1.70) and a marginally significant effect of math anxiety (β = − 0.253, t = 1.95, P = 0.053, BF10 = 2.94), but their interaction was not significant (β = 0.018, t = 0.14, P = 0.888, BF10 = 0.61). Accordingly, we rejected the moderation model.

Mediation analysis: gender did not significantly predict math anxiety (β = − 0.054, t = − 0.56, P = 0.579, BF10 = 0.23), and no further mediation analysis was conducted.

To conclude, both gender and math anxiety showed a direct effect on career choice math intensity for high school students. Female students tended to choose careers with lower math intensity, and both male and female students with high math anxiety tended to choose careers with lower math intensity. Importantly, no moderating or mediating effects were evident (see Fig. 2).

Fig. 2

Links between gender, math anxiety, and math intensity of career choice for high school students. Illustration of gender and math anxiety as predictors of career choice math intensity for high school students. A Both gender (β = 0.187, t = 2.00, P = 0.047, BF10 = 2.35) and math anxiety (β = − 0.240, t = − 2.58, P = 0.011, BF10 = 5.35) significantly predicted career choice [F_{(2, 104)} = 5.63, P = 0.005, r² = 0.098]. B Gender did not significantly predict math anxiety (β = − 0.054, t = − 0.56, P = 0.579, BF10 = 0.23)

University students

Moderation analysis: math anxiety (β = − 0.235, t = 2.31, P = 0.023, BF10 = 3.26), but not gender (β = 0.040, t = 0.39, P = 0.693, BF10 = 0.42), significantly predicted career choice math intensity [F_{(2, 97)} = 3.19, P = 0.046, r² = 0.062]. The main effect of math anxiety on career choice (β = 0.306, t = 2.32, P = 0.022, BF10 = 1.00) remained significant in moderation analysis, with non-significant main effects of gender (β = 0.046, t = 0.45, P = 0.652, BF10 = 0.34) and gender and math anxiety interaction (β = 0.112, t = 0.85, P = 0.395, BF10 = 0.87). The moderation model was non-significant [F_{(3, 96)} = 2.63, P = 0.076, r² = 0.069] and thus was rejected.

Mediation analysis: gender was a significant predictor of math anxiety (β = − 0.260, t = − 2.67, P = 0.009, BF10 = 4.65), so we conducted mediation analysis. Gender did not predict career choice math intensity (β = 0.101, t = 1.01, P = 0.316, BF10 = 3.26), but math anxiety did (β = − 0.245, t = − 2.50, P = 0.014, BF10 = 3.26). Mediation analysis (see Fig. 3 for illustration) indicated the existence of a significant indirect effect of gender on career choice math intensity via math anxiety (95% CI 0.011 to 0.146; P = 0.012, as tested by a bias-corrected bootstrap procedure). Note that a direct effect of the independent variable on the dependent variable is not mandatory for mediation analysis (Hayes, 2013).

Fig. 3

Links between gender, math anxiety, and math intensity of career choice for university students. Among university students, gender was a significant predictor of math anxiety (β = − 0.260, t = − 2.67, P = 0.009, BF10 = 4.65), but did not predict career choice math intensity (β = 101, t = 1.01, P = 0.316, BF10 = 3.26), while math anxiety did (β = − 0.245, t = − 2.50, P = 0.014, BF10 = 3.26). Mediation analysis indicated a significant indirect effect of gender on career choice math intensity via math anxiety (95% CI 0.011 to 0.146; P = 0.012, as tested by a bias-corrected bootstrap procedure) [F_{(2, 97)} = 3.19, P = 0.046, r² = 0.062]

We conclude that for university students, gender had an indirect effect on career choice math intensity via math anxiety. Females reported higher math anxiety levels, and these were related to a choice of a career with a lower math intensity.

Discussion

This study investigated the links between gender, math anxiety, and math intensity of career choice at different educational phases. We found that gender, but not math anxiety, predicted career choice math intensity for primary school students; female students were more interested in careers with lower math intensity than male students. High school students showed different trends, as both gender and math anxiety directly predicted career choice math intensity, without interactions between them. In high school, female students and students with high math anxiety (both genders) were more interested in careers with a lower math intensity. For university students, the association of gender on career choice math intensity was indirect and mediated by math anxiety. Unlike the preceding educational phases, female university students reported higher math anxiety levels than males, and this predicted the choice of a career with a lower math intensity. Therefore, our preliminary results suggest that the influence of math anxiety on career choice may emerge in secondary school (Eidlin Levy et al., 2021; Huang et al., 2019; Megreya & Al-Emadi, 2023) and persists into higher education (Ahmed, 2018; Daker et al., 2021; Morán-Soto & González-Peña, 2022).

Gender differences in associations between math anxiety and math intensity of a selected career

Our findings suggest that the interplay between gender and math anxiety as predictors of variance in career choice seems to change across development. For primary school students, gender but not math anxiety predicted preferences for a math-intensive career. Very young students can experience math anxiety (Gunderson et al., 2018; Primi et al., 2020; Tomasetto et al., 2021). However, the influence of math anxiety on later outcomes, such as career choice, may be insignificant at this educational phase, as students do not need to set career goals (Kohen & Nitzan, 2021; Reinhold et al., 2018). Even so, primary school students seemed to have already internalized gender-biased stereotypes, which influenced their career plans (Moè, 2018; Teliousi et al., 2020).

Both gender and math anxiety made unique contributions to career choice for high school students, with no interaction between them. Consistent with the assumption that math anxiety increases over school years (Devine et al., 2018; Hart & Ganley, 2019; OECD, 2013) and becomes differentiated from other anxieties (Carey et al., 2017), we found that the link between math anxiety and possible outcomes, such as career choice, appeared among high school students. High anxiety has been associated with avoidance behavior (Dymond & Roche, 2009) and hence possible low academic achievement, and as indicated in our results, enrollment in orientation studies with lower math intensity. Career goal setting in high school, such as enrollment in math-intensive orientation studies and advanced math courses, can impact later career paths. Avoidance behavior is related to lower math competence in the long term (Daker et al., 2021), and this may be a predictor of Delaney and Devereux (2019), Tandrayen-Ragoobur and Gokulsing (2022) or even a prerequisite (Ayalon & Yogev, 2005) for acceptance into math-intensive studies in higher education. Past research suggests that math anxiety levels may be higher for high school students (OECD, 2013) than for university students (Hart & Ganley, 2019), suggesting that extreme math anxiety levels may even prevent students from attending higher education.

Interestingly, the association of math anxiety and career choice was independent from gender, and gender did not predict math anxiety for high school students. The findings suggest that math anxiety has a negative link with students’ aspirations for a career with a high math intensity, regardless of gender (as in Ahmed, 2018; Daker et al., 2021). Continuous negative math experiences resulting in avoidance patterns (Choe et al., 2019; John et al., 2020) may account for these findings, challenging gender segregation theories (Else-Quest et al., 2010; Stoet et al., 2016).

For university students, the effect of gender on career choice math intensity was not direct, but mediated by math anxiety. Female students reported higher math anxiety levels, and these were associated with preferences for careers with a lower math intensity. The findings also indicated a negative association between math anxiety and career choice math intensity among university students. However, the characteristics of the university sample may account for the differences between high school and university samples, as not all high school graduates continue on to higher education. Furthermore, as in other countries, female Israeli students are more likely than male students to enroll in higher education, but are less likely to enter scientific fields (OECD, ). Social–behavioral aspects, such as ethnicity or behavioral disorders, can affect drop out (Peguero et al., 2019, 2021). Personal–social resilience factors, such as social support (Tandrayen-Ragoobur & Gokulsing, 2022), positive contact with instructors and peers (Meyer & Strauß, 2019), or personal traits (Isphording & Qendrai, 2019) can account for choosing and persisting in math-intensive studies, especially among females. Therefore, we can speculate that females entering higher education and specifically taking STEM courses may have had social–personal resilience factors and thus enrolled in math-intensive studies despite high math anxiety levels.

Associations between math anxiety and math intensity of a chosen career: a developmental perspective

Although we did not directly compare different educational phases, the preliminary findings suggest some interesting trends in the associations between math anxiety and math intensity of a chosen career across development. The link between math anxiety and career choice was evident for older but not younger students. This pattern may result from the more restrictive structure of the Israeli educational system for older students, who are requested to choose a field of study and set more specific career goals and plans (Brown & Lent, 2019; Kohen & Nitzan, 2021). Students’ career aspirations are also adjusted by their personal experiences along the educational trajectory (Reinhold et al., 2018). Accordingly, the findings could represent the cumulative effect of math anxiety on the choice of a career. Students with high math anxiety may avoid participating in math activities, such as doing homework, on the micro level; this, in turn, can harm their math grades and decrease their interest in math-intensive careers on the macro level (Daker et al., 2021; Soysal et al., 2022).

Suggestions for future research

This study, along with other recent studies (Ahmed, 2018; Daker et al., 2021; Eidlin Levy et al., 2021; Huang et al., 2019; Megreya & Al-Emadi, 2023; Morán-Soto & González-Peña, 2022), indicates math anxiety may affect students' decisions to enroll in math-intensive courses and follow math-intensive careers. However, further research is required to address some important issues. It is not clear whether the association between math anxiety and avoidance of math-intensive studies is exclusive to females (Eidlin Levy et al., 2021; Huang et al., 2019; Megreya & Al-Emadi, 2023; Morán-Soto & González-Peña, 2022) or evident across genders, as we and others have found (Ahmed, 2018; Daker et al., 2021). Replication studies controlling for possible covariates such as trait or test anxieties (Carey et al., 2017; Devine et al., 2012; Hill et al., 2016) and including measures of math competence would address this question. Furthermore, qualitative research, as well as investigation of the relations between parents’ and teachers’ math anxiety and students' career plans, may reveal interesting aspects regarding the social influences of career choice. As our research was cross-sectional, further longitudinal research is essential to probe the early influences of math anxiety on academic achievements and career choices. Studies should also compare students with different academic experiences, such as freshmen vs. upperclassmen. Finally, future research may assess both enrollment in math-intensive studies and grades in these studies to capture the influence of math anxiety on academic and career behavior on both macro- and micro-levels, as suggested by Daker et. al. (2021).

Conclusion

In our study, math anxiety directly predicted career choices for high school and university but not primary school students. Gender had a direct effect for younger students, as female students attending primary and high school preferred careers with a lower math intensity. The effect of gender on career choice for university students was not direct but mediated by math anxiety. Thus, it is crucial to identify young students with math anxiety and provide appropriate math anxiety reduction programs (Furner & Duffy, 2022; as suggested by Rozek et al., 2019) to reduce the possible cumulative effect of math anxiety on academic achievement and career choice.