Do Girls Really Experience More Anxiety in Mathematics? Flashcards

Thomas Goetz, Madeleine Bieg, Oliver Lüdtke, Reinhard Pekrun, and Nathan C. Hall University of Koblenz, 2013

1
Q

What did previous study find regarding gender differences in Math and anxiety?

What are the implications?

A
  • females typically obtain similiar or (only) slightly lower results in math compared to males
  • females report higher levels in anxiety in math (meta-analyses)
    • math anxiety negatively predicts course enrollment, career choices and learn
    • hence, it may have contributed to the underrepresentation of females in STEM fields
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What may be limitations in previous findings about math anxiety and gender?

A
  • previous studies used self-reports of traitlike (habitual) anxiety
  • this can lead to very results compared to state anxiety (in real-life situations)
  • this study wants to assess if there is an actual state anxiety difference between men and women
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What are the theoretical differences between state and trait anxiety?

A
  • empirical findings suggest that trait anxiety (self-reported) are influenced by subjective beliefs
  • these beliefs are less likely to influence state anxiety
  • this is an assumption of the accessibility model of emotional self-reports:
    • state measures are assumed to evaluate emotions
    • trait measures are assumed to rather evaluate belief about emotions
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

How may subjective beliefs of competence influence math anxiety?

A
  • competence beliefs may play a central role in self-reports of trait emotions
    • it may be an “antecedent of anxiety
  • girls typically report lower perceived competence, self-efficacy and perceived ability
    • this may be due to gender stereotypes as e.g. “Girls and math is a bad fit”
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What is the hypothesis by the current study from Goetz et al. (2013)?

A

They propose that:

  • the gender gap in math anxiety is due to using self-reports methods that evaluate trait anxiety
    • this trait anxiety may be influenced by personal comppetence beliefs
    • measures of state anxiety during math exercises and test would show a smaller gender gap and be less influenced by competence beliefs
    • even though not a focus: that girls and boys get relatively similar math grades
  • previous studies found this difference in trait and state measures (e.g. coping strategies) but not on math anxiety
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What were the two studies by Goetz et al. (2013) and which variables were assessed?

A
  • Both studies evaluated trait and state anxiety, self-reports of perceived competency and math grades
  • varrious age groups as well as test and class-related anxiety was assessed
  • both studies used experience-sampling methods
  • Study 1
    • 5th to 10th graders
    • test anxiety was assessed with trait and state anxiety measures
  • Study 2
    • 8th and 11th graders
    • class-related anxiety was assessed with trait and state anxiety measures
      • state measures during regular class
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What is tth experience-sampling method?

A

The experience sampling method (ESM) is a strategy for gathering information from individuals about their experience of daily life as it occurs. It is a phenomenological approach, meaning that the individual’s own thoughts, perceptions of events, and allocation of attention are the primary objects of study.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Sample of study 1:

A
  • multiple grade levels of the top track of the education system in Germany (ca. 1/3 of student cohort)
  • 584 students (24 classes, six schools)
  • grades 5 to 10
  • 45% female, 55% male
  • mean age: 13.67, SD: 1.84
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Sample of study 2:

A
  • 111 students (2 to 4 students selected randomly from 41 classrooms across seven schools)
  • grades 8 and 11
  • 53% female, 47% male
  • mean age: 15.96, SD 1.71
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Procedure of Study 1 and 2, respectively.

A

Both studies: standardized questionnaire at the beginning followed by state self-report measurements

Study 1:

  • state math test anxiety was assessed immediately before the test and twice during it ( after 1/3 and after after 2/3 of the test)
  • the self-report meaasures were integrated into the answer sheet

Study 2:

  • class-related anxiety was assessed via a digital questionnaire presented on a personal digital assistant (PDA) following a randomised audible signal
    • once per class
    • over two weeks
      • ca. five assessments per student
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

How were the variables assessed?

  1. Anxiety
A
  1. Anxiety:
  • in study 1:
    • trait math anxiety was assessed using the Achievement Emotions Questionnaire (four items, how students typically felt during a math test)
    • state math anxiety was measured with 1 item:
      • “I am anxious”
    • both used a five-point Likert-scale
  • in study 2:
    • one item, trait anxiety: “How much anxiety do you generally experience during math class”?
    • one item, state anxiety: “How much anxiety do you experience duriing this class?”
    • both had a five-point Likert scale
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

How were the variables assessed?

  • perceived competence:
A
  • operationalised as self-efficacy and self-concept beliefs
  • study 1:
    • self-efficacy was measured with a four item scale from PISA assessments
    • e.g. “I am confident that I can understand even the most difficult content in math”
    • 5-point Likert scale
  • study 2:
    • academic self-concept measured with three items of the Self-Description Questionnaire
    • e.g. “math is one of my best subjects”
    • 5-point Likert scale
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

How were the variables assessed?

  • Achievement
A

In both studies:

  • operationalised as students mid-term grades
    • typically based on one single written exam and oral exams
  • grades range from 1 (very good) to 6 (insufficient)
    • for this studies values were inverted to facilitate comprehension (hence, 6 = very good, 1 = insufficient)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

How was the data analysed and modeled?

A
  • multileve, intraindividual modeling approach to account for the nested structure of the data
  • both studies used Hierarchical Linear Modeling software
  • multilevel analysis consisted of three levels:
    1. measures within students
    2. student level
    3. class level
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Explain the implications of levels one:

  • measures within students
A

outcome variable were anxiety scores.

sudy 1

  • trait anxiety: anxiety score divided by number of items
  • state anxiety: three measurements, one before and two during test

Study 2:

  • trait anxiety: outcome on single item
  • experience sampling method, four ratings per student (on average)
  • variables were coded 0 (state anxiety and 1 (trait anxiety)
  • therefore, intercept: Y000 describes overall mean state anxiety for males (because male/female was also coded 0 and 1)
  • hence, Y100 indicates the difference between trait and state anxiety…positive scores indicate that trait scores were larger
  • This variable’s effect (γ100) can be interpreted
    as the difference between trait and state anxiety
    scores, with positive values indicating that trait scores
    were higher than state scores
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Explain the implications of level two:

  • student level
A

two level two variables + interaction:

  • gender (0 = male, 1 = female; Y010, uncentered)
  • study 1: self efficacy, study 2: self-concept; Y020 z scores standardised across persons
  • Gender * Competence (Y030, multiplicative term)
17
Q

Explain level three:

  • class level
A
  • classes in which students were nested
    • in order to take into account clustering of students within classes for estimating SE
18
Q

What are the cross level interactions in this study?

  • Level 1 - Level 2
A

three cross-level multiplicative terms:

  • Trait/State * Gender (Y110)
  • trait/state * Competence (Y120)
  • trait/state * Gender * Competence (Y130)

these terms indicate effects of gender, competence and the combined effect of gender and comptetence on the difference between trait and state anxiety

19
Q

Which models were used to test the hypotheses?

A

each model was set up as a slopes-as-outcome model

  1. model 1 examined the effects of gender on the difference between state and trait anxiety (Y110)
    • trait/state anxiety * Gender interaction
  2. model 2 examined the effects of competence on the difference between state and trait anxiety (Y120)
    • trait/state anxiet * competence interaction
  3. model 3 eamined if the effect of gender on trait - state anxiety differences were smaller when competence beliefs were included as a mediator of gender effects (Y110, Y120)
  4. model 4 included the three-way interaction of trait/state anxiety, gender and competence (Y130)

in all models main effects were included (Y010, Y020, Y030)

20
Q

Preliminary analysis:

A
  • both studies: grirls reported higher trait anxiety
  • and lower competence beliefs
    • these effectsizes were medium to large
  • girls and boys did not significantly differ on math achievement and state anxiety
21
Q

Main analyses (table)

A
22
Q

Results for model 1

  • and main affects of gender and state/trait anxiety
  • testng if gender predicted the difference in trait and state anxiety (interaction terms)
  • implications
A

main effects:

  • main effect of type of measure (trait/state variable, Y100) was significant in study 1 but not study 2
  • main effects gender on anxiety (Y010): not significant in either study

Interaction effects:

  • trait/state anxiety * gender (Y110) significant in both studies

implications:

  • supports hypothesis 1 thhat gender informs differences between self-reported trait and state math-anxiety
    • discrepancy between state and trait anxiety higher for grirls
23
Q

Findings of model 2:

  • interaction of trait/state * competence (Y120)
A
  • was significantly negative in both studies
    • higher competence beliefs corresponded with smaller differences in trait and state anxiety
24
Q

Findings of model 3.

  • trait/state * competence (Y120)
A
  • trait/state * competence interaction remained significant
  • trait/state * gender interaction was significant in study 2 but not in study 1
  • the general effect of gender on trait-state discrepancy was smaller in both studies due to including of a trait/state interaction term
25
Q

findings of model 4

  • adding the trait/state * gender * competence interactionterm
A
  • effects of gender on trait/state discrepancy (Y110) and competence on trait-state discrepancy (Y120) were additive - because Y130 did not reach significance in either study
26
Q

interpretation of findings from model 1 to 4:

A
  • effect of gender on trait-state discrepancy can be interpreted as a moderator effect:
    • gender played a significant role in informing the size of the disccrepancy between state and trai anxiety
    • this moderating effect was partly mediated by perceived competence
      • this was evidenced by the reduction in the state/trait*gender coefficient (Y110) whetn trait/state anxiety* competence was added
    • this support H2 that girls’ trait-state discrepancy was associated with lower levels of perceived competence
27
Q

All in all, what are the findings of the study (also relative to prior research?

A
  • supported the well known gender gap in trait math anxiety (self-reported habitual anxiety)
  • however, that girls do not report higher state math anxiety than boys
    • this means that they actually don’t experience more anxiety than boys when doing or being tested on math
  • lastly, that lower math competence beliefs in grils may be partly responsible for higher trait anxiety in girls
28
Q

How do these findings fit into the accessibility model of emotional self-reports?

A

In the accessibility model of emotional self-reports:

  • state measures are thought to evaluate someone’s emotion
  • trait measures are understood to reflect individual’s beliefs about emotions
  • in this study competence beliefs may influence cognitive appraisals about emotions (power: 37.9% and 54.6% in studies 1 and 2, respectively for explaining discrepancy in state-trait anxiety)
  • there are also other math related cognitions that might have had an influence, e.g.:
    • perceived value
    • content difficulty
    • achievement expectations
    • etc.
  • also gender stereotypes may have influenced these cognitive appraisals/ competence beliefs and therefore trait anxiety
  • that reflective cognitive processes lead to an self-overestimation of girl’s anxiety may be further supported by the fact that yhere were no gender differences in math achievement
29
Q

Why does it matter how and why girls overestimate their habitual math anxiety?

A
  • because perceived trait anxiety can negatively effect subjective well-being, motivation and learning behaviour
  • can also impact decision making processes and hence,
    • beliefs about math anxiety may be a reason for underrepresentation of females in math-intense domains (e.g. STEM)
  • this study constsited of students of the highest educational track in GErmany, hence a large proportion of them are high-achievers and assumed to take positions of leadership later
    • but even of these female high achievers, some may not choose math-intense domains simply due to lower competence beliefs and trait math anxiety
30
Q

What are some poritive implications of this study’s findings?

A
  • this trait math anxiety may be addresable via improving girl’s self-defeating cognitions and emotions in math
    • e.g. by explicitly informing them that achievement and anxiety in math does not sig. differ from those of the boys
    • cognitive interventions might be usable to reduce the gender gap in trait math anxiety
  • if successful this might have positive ecomic effects for society via addressing the international shortage of STEM- related workers
    • (math-intense domains)