Math and Gender

1
Math and Gender

• Luigi Guiso
• Ferdinando Monte
• Paola Sapienza
• Luigi Zingales

2
Motivation

• There are well-established gender differences in
  math and reading test performance. What is the
  cause?
• Environment
• Biology
• Strongest argument for biology is the existence
  of some gender differences in cognitive abilities
• Men better at
• aiming
• spatial ability
• Men worse at
• verbal fluency and recall
• These cognitive abilities linked to biological
  differences between gender.
• If they can be linked to math and reading
  abilities ? biology argument.

3
Recent revival

• Debate traditionally intense why so few women in
  top science departments? MIT
• Only 8 are women in science (Biology, Physics,
  Mathematics etc.)
• Only one out 38 professors in the Math
  department! (Gigliola Staffilani)
• Debate recently revived by Larry Summer, ex
  Harvard President, who ventured to argue that
  from a pure scientific point of view one cannot
  exclude there is a biological component
• Because of this he lost his job as Harvard
  President
• Because of this his appointment as Obamas lead
  economic advisor has been heavily criticized

4
Facts actors
Larry Summers
5
Approach

• Cognitive differences have been found in all the
  populations (except the Inuit or Yupik )
• But environmental (cultural) differences across
  countries are huge
• Use a large sample of comparable data across
  countries with different attitudes toward women
  to determine how much of the difference in
  performance is environmental

6
PISA (Program for International Student
Assessment)

• 276.000 students in 41 countries tested at age 15
• In 2003, 4 tests
• math, problem solving, science, reading
• Lots of data on
• Intrinsic motivation (taste - driven)
• Extrinsic motivation (instrument driven)
• Stress levels
• Tests are culture free

7
Math tests

• Scores reflect ability to apply mathematics in
  solving real-life problems
• Questions in math cover
• space and shape (geometry)
• change and relationship (algebra)
• quantity (arithmetic)
• uncertainty (probability)
• in a range of difficulty that goes from the need
  of simple mathematical operations to complex
  thinking.
• Math scores scaled to have mean of 500 and
  standard deviation of 100 in the OECD students
  population.

8
(No Transcript)
9
Gender Gap in Math
10
Gender Gap in Reading
11
Focus

• Focus so far within countries
• At this level a gender gap in math (almost) in
  all countries
• But there are marked differences in the size of
  these gaps across countries.
• Why?
• They have been overlooked
• Explaining them is our focus

12
Gender Gap in Math by Country
ITALY
13
Measures of Women Emancipation

• Gender gap index from the Global Competitiveness
  Report (WEF, 2006)
• World Value Survey
• percentage of people that "disagree" with
  assertions like "When jobs are scarce, men should
  have more right to a job than women".
• Participation to the labor force (UNESCO)
• Female-to-male ratio of tertiary enrollment
  (UNESCO)

14
Women emancipation index by Country
SWEDEN
ITALY
TURKEY
15
Math Gender Gap and GGI
16
Math Gender Gap and Women Participation
17
It is not just economic development

• We run the regression at the individual level
• Insert country dummies (that control for all the
  possible institutional differences)
• Insert the interaction between gender and GGI
• The interaction is positive and statistically
  significant effect robust to other
  institutional differences

18
Effect economic sizeable

• Raising Turkey women emancipation to the level
  prevailing in Sweden would close the math gender
  gap!
• Interestingly, increased women emancipation not
  only improves the math gap but also strengthens
  women advantage in reading
• Womens performance improves across the board
• Men performance is no worse
• What is unaffected is the within gender relative
  performance
• Women do relatively better in reading than in
  math and men vice versa, independently of
  societys women emancipation

19
How does women emancipation affect scores ?

• 1) Economic channel Higher payoff - higher
  investment
• more hours in homework and classes
• more effort in each class
• 2) Psychological channel -
• More self confidence
• Less anxiety

20
How does women emancipation affect scores ?

• 3) Educational channel
• Teaching style
• Discipline
• Different approach to subjects
• 4) Sociological channel
• Role model
• Peer pressure

21
1) Economic channel

• Does women emancipation increase
• Hours spent by women in math courses? NO
• Hours spent by women in math homework? NO
• Effort put by women in studying math
• (measured as the marginal effect of an extra hour
  of class)? NO

22
2) Psychological channel

• Does women emancipation increase
• Women intrinsic motivation?
• Women extrinsic motivation?
• Women self-confidence?
• Or decrease
• Women level of anxiety?

23
Variables

• Self assessments (To what extent do you agree
  with a bunch of statements) of
• Intrinsic motivation
• Extrinsic motivation
• Self confidence 1 (self concept)
• Self confidence 2 (self efficacy)
• Anxiety

24
Variables definitions

• Intrinsic motivation
• I enjoy reading about mathematics. ()
• I look forward to my mathematics lessons. ()
• I do mathematics because I enjoy it. ()
• I am interested in the things I learn in
  mathematics. ()
• Extrinsic motivation
• Making an effort in mathematics is worth it
  because it will help me in the work that I want
  to do later on. ()
• Learning mathematics is worthwhile for me because
  it will improve my career .
  ()
• Mathematics is an important subject for me
  because I need it for what I want to study later
  on. ()
• I will learn many things in mathematics that will
  help me get a job. ()

25
Variable definitions

• Self efficacy
• How confident do you feel about having to do the
  following calculations?
• Self-concept
• I am just not good at mathematics.
• I get good in mathematics. ()
• I learn mathematics quickly. ()
• I have always believed that mathematics is one of
  my best subjects. ()
• In my mathematics class, I understand even the
  most difficult work. ()

26
Variable definitions

• Anxiety
• I often worry that it will be difficult for me in
  mathematics classes. ()
• I get very tense when I have to do mathematics
  homework. ()
• I get very nervous doing mathematics problems.
  ()
• I feel helpless when doing a mathematics problem.
  ()
• I worry that I will get poor in
  mathematics. ()

27
Female-Male Gap
28
Results

• Motivation and anxiety matter
• But no evidence that women emancipation works
  through an increase in intrinsic or extrinsic
  motivation, an increase in self confidence, or a
  reduction in anxiety
• In fact, where women are more emancipated they
  have
• lower relative self concept in math
• higher math anxiety

29
3) Educational channel

• Discipline
• Correlation between women emancipation and
  discipline? No correlation
• Different approaches to subjects (more emphasis
  in math)
• Correlation between women emancipation and
  importance of math? No correlation
• Differences in teaching style
• Foster different learning environments? No
  correlation

30
4) Sociological channel

• We compute the average math score of the other
  boys and the other girls in the same school.
• We run a micro level regression of math scores on
  these variables (level and interacted with
  gender) for each country
• Estimate reflects the importance in that country
  of the role model (or peer effect)

31
Pure differential peer effect and GGI
32
Results

• Pure differential peer effect less important in
  countries with higher GGI
• Consistent with the idea that role models are
  different in more emancipated countries
• in countries with more women emancipation,
  women performance in math less (positively)
  affected by the performance of other girls and
  less (negatively) affected by that of other boys
  

33
Looking forward

• Very recently Steven Levitt has looked at this
  issue again. He finds that
• Using US panel data, a gender gap emerges early
  at school kids perform initially equal but a
  gap emerges as they grow older
• Confirms our findings in a different dataset when
  he uses the same countries
• But correlation with women emancipation
  disappears when Muslin countries are added!
• Why?

34
Looking forward

• In Muslin countries male and female go to
  same-sex schools
• girls do not lag boys in countries with same-sex
  schooling, even if in the countries where women
  are much less emancipated
• Not exposing them to men seems to be enough to
  avoid the effect of culture on gender gap in math

35
Conclusions

• We identify a strong cultural factor in women
  test performance
• Where women are treated more equally, they
  exhibit a stronger absolute advantage in reading
  and a weaker absolute disadvantage in math.
• This positive effect does not work through
• Standard economic incentives
• Psychological effects
• Different educational styles
• Most plausible channel seems a role model effect