Gender and Educational Attainment in Schools

Stephen Machin and Sandra McNally

Girls doing much better!

• Headline figure for 2002: 9% gap in attainment at GCSE (5+ A*-Cs)

• Girls outperforming boys in every subject at GCSE

• At all stages of education (for all ethnic groups)

Explanations: media

• ‘Laddish culture’/ ‘macho’ peer groups / ‘too much football’

• ‘Teachers giving up on boys’

• ‘Extinction’ of male teachers at primary school

• Girls mature earlier

• Coursework element in exams

Some research questions• At what stage in education is this gap most

important? • How is gender gap related to changes at

school/exam system or wider social and economic changes, e.g. higher education and labour market participation of women; decline in male teachers; cultural changes

• Can policy make a difference?• How does gap in school attainment affect

differences in post-compulsory schooling, labour market outcomes?

What is going on? – Data sources

• National Key Stage data-sets: children tested at 7, 11, 14 and 16 (GCSE).

• General Household Survey: annual survey of 9,000 households from 1972

• Longitudinal data-sets: of all children born in a particular week in March 1958 (National Child Development Study) and April 1970 (British Cohort Study)

Comparing cohort and national data

1975 national data and the NCDS All Boys Girls 5+ O-Levels A-C National data 0.23 0.22 0.23 NCDS 0.18 0.17 0.18 English O-Level A-C for entrants National data 0.55 0.66 NCDS 0.69 0.78 Maths O-Level A-C for entrants National data 0.59 0.53 NCDS 0.76 0.70 1986 national data and the BCS All Boys Girls 5+ O-Levels A-C National data 0.27 0.26 0.27 BCS 0.31 0.32 0.31 English O-Level A-C for all National data: 0.41 0.35 0.46 BCS 0.43 0.39 0.46 Maths O-Level A-C for all National data: % leavers 0.32 0.35 0.30 BCS 0.33 0.37 0.30

Gender differences at end of Primary school NCDS Age 11 in

1969 BCS Age 10 in

1980 KS2

Age 10/ 11 in 2002 Average regressions Maths .83 (.49) 2.73 (.53) 2.48 (.07) Reading -.3 (.49) -4.03 (.51) -4.14 (.07) Quantile regressions

10th percentile Maths -3 (1.4) 0 (.58) 0 (.59) Reading 0 (1.42) -4 (.59) -3 (.57) 50th percentile Maths 3 (2.98) 3 (2.98) 4 (1.49) Reading 0 (5.87) -3 (1.5) -7 (3.44) 90th percentile Maths 2 (1.15) 3 (.87) 5 (.58) Reading 3 (1.66) -3 (.86) -3 (1.45) Notes: Coefficient on dummy for boys reported; Standard errors in parentheses; Sample sizes are NCDS Maths 14129, NCDS Reading 14133, BCS Maths 11706, BCS Reading 12775, KS2 1996 Maths 539074, KS2 1996 Reading 535711, KS2 1998 Maths 539074, KS2 1998 Reading 574152, KS2 2002 Maths 611396, KS2 Reading 608674.

Raw gender differential over time: secondary schools

1+A*-C 5+ A*-C

Maths A*-C

English A*-C

(1) 1974 NCDS -0.059 -0.018 0.042 -0.090 (0.008) (0.006) (0.006) (0.007) (2) 1986 BCS -0.030 0.009 0.068 -0.072 (0.012) (0.011) (0.011) (0.012) (3) 1998 KS Data -0.094 -0.097 -0.010 -0.035 (0.001) (0.001) (0.001) (0.001) (4) 2002 KS Data -0.089 -0.111 -0.034 -0.060 (0.001) (0.001) (0.001) (0.001)

Notes: Marginal effect on dummy for boys reported; Standard errors in parentheses; Sample sizes are (1) 14311 (cols 1-4) 11987 (cols 5-6); (2) 6949 (cols 1-4); 5621 (col.6); (3) 551,046 (cols 1-2); 528780; 491488; (4) 586356 (cols 1-2) 571301; 527051.

Gender differential in secondary school: added value model

1+A*-C 5+A*-C Maths A*-C

English A*-C

(1) 1974 NCDS -0.08 -0.021 0.019 -0.113 (0.009) (0.004) (0.003) (0.007) (2) 1986 BCS -0.045 -0.008 0.053 -0.098 (0.013) (0.011) (0.012) (0.013) (3) 2002 KS Data -0.083 -0.137 -0.049 -0.087 (0.001) (0.002) (0.002) (0.002)

Notes: Marginal effect on dummy for boys reported; Standard errors in parentheses; Sample sizes are (1) 14311 (cols 1-4) 11987 (cols 5-6); (2) 6949 (cols 1-4); 5621 (col.6); (3) 586356 (cols 1-2) 571301; 527051.Controls for reading score, age 11, maths score, age 11, missing variable dummies for Maths and English

Implications

• Gaps appear to have widened out in secondary school years over time.

• English: wide gap in favour of girls that persists over time.

• Maths: gender gap changes over time (in favour of boys in 70s and 80s, in favour of girls in 90s and 00s)

Economic framework

• Education of children viewed as family investment (Becker model..)

• Children begin life with ‘inherited’ ability. Parents make investments….influenced by preferences, income and fertility.

• Children’s ability and level of parental income and home investment determine schooling attained by children

Higher education of mothers

• Effect through income – higher income, more education

• Effect on home investment –quantity/quality

• Tastes/attitude to education of children

Most research suggests that mother’s education is more influential than father’s; and has a bigger influence on girl’s education

Education of mothers, by survey year

0

0.1

0.2

0.3

0.4

0.5

0.6

1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1998 2000

Higher education Other education

Link between income and the gender differential?

• Relationship between income and education has changed over time – much stronger now

• Has this benefited girls more than boys? – depends on mechanism

- At least one study showing that increases in household income lead to greater investment in girls’ schooling (Glick and Sahn, 2000) – but developing country context.

Does Parental Education and Income Explain Gender Gaps in Age 16 Attainment?

5+ A-C O Levels (Coefficient, SE, on Male)

NCDS Age 16 BCS70 Age16 (1) Basic Value Added Specification

-.021 (.004)

-.008 (.011)

(2) (1) Plus Parents’ Education

-.020 (.004)

-.011 (.011)

(3) (1) Plus Parental Income -.018 (.004)

-.008 (.014)

(4) (1) Plus Parents’ Education and Parental Income

-.018 (.004)

-.010 (.014)

Notes: Sample sizes range from 9025 in (4) to 14257 in (1) for NCDS and from 4846 in (4) to 7307 in (1) for BCS70. The same pattern of results emerge if restricted to the smallest sample size within each cohort. All specifications include age 10/11 maths and reading test scores.

Does parental education and income explain gender gap at age 16? Evidence from the GHS

Controls None Parental

Education Earnings Parental

education & Earnings

Surveys

Obs Coef S.E. Coef S.E. Coef S.E. Coef S.E.

77-79 1467 -0.002 [0.021] -0.007 [0.018] -0.004 [0.021] -0.007 [0.018] 78-80 1506 0 [0.021] -0.002 [0.020] -0.001 [0.021] 0 [0.020] 79-81 1593 0.007 [0.020] 0 [0.020] 0.005 [0.021] 0.003 [0.020] 80-82 1516 0.026 [0.021] 0.024 [0.021] 0.026 [0.021] 0.031 [0.021] 81-83 1396 0.029 [0.022] 0.03 [0.022] 0.029 [0.022] 0.035 [0.022] 82-84 1221 0.026 [0.024] 0.025 [0.024] 0.029 [0.024] 0.031 [0.024] 83-85 1148 0.001 [0.025] 0.002 [0.025] 0.002 [0.025] 0.005 [0.025] 84-86 1129 -0.004 [0.025] -0.005 [0.025] -0.003 [0.025] -0.002 [0.025] 85-87 1097 -0.028 [0.026] -0.031 [0.026] -0.028 [0.026] -0.032 [0.026] 86-88 975 -0.042 [0.026] -0.038 [0.026] -0.038 [0.026] -0.038 [0.026] 87-89 857 -0.067 [0.026] -0.063 [0.026] -0.068 [0.026] -0.064 [0.026] 88-90 771 -0.05 [0.026] -0.041 [0.026] -0.044 [0.026] -0.038 [0.026] 89-91 746 -0.066 [0.027] -0.066 [0.027] -0.061 [0.027] -0.063 [0.027] 90-92 718 -0.071 [0.028] -0.062 [0.028] -0.055 [0.028] -0.055 [0.028] 91-93 666 -0.12 [0.029] -0.106 [0.030] -0.114 [0.030] -0.105 [0.030] 92-94 696 -0.089 [0.032] -0.077 [0.033] -0.097 [0.032] -0.078 [0.033] 93-95 709 -0.124 [0.034] -0.126 [0.036] -0.135 [0.035] -0.128 [0.036] 94-96 790 -0.095 [0.035] -0.116 [0.037] -0.111 [0.036] -0.121 [0.037

Research on male/female characteristics with impact on educational attainment

• Attributes: Girls: more attentive; longer concentration span; give fewer discipline problems; but less confident than boys

• Learning styles: ‘Boys show greater adaptability to more traditional approaches to learning…memorising abstract, unambiguous facts that have to be acquired quickly…willing to sacrifice deep understanding for correct answers achieved at speed’

Coursework vs. exams - Study of GCSE Maths (Stobart, 1992): Boys

achieve a small mark advantage in exams, which offset girls’ small advantage in coursework.

- Creswell (1990): 1989 GCSE exams in English, Maths and Science. Girls’ marks for coursework higher; Maths and Science: Boys’ marks for exams higher.

- Ranking of subjects in terms of coursework (Stobart et al. 1992), Improved performance of girls directly related to the weighting and type of coursework

Coursework vs. exams

• Elwood (1995): study of GCSE Maths and English. Gender differential present in courses when coursework element was reduced; gender gap most pronounced for syllabuses not 100% assessed by coursework.

• In English, restriction imposed in 1994 from 100% to 40% coursework - performance gap did not reduce.

Influence of method of assessment

Multiple choice assessment: The Third International Mathematics and Science Study (TIMSS) – boys doing considerably better than girls in Maths and Science, whereas doing about the same in GCSE (Key, Harris and Fernandes, 1996)

The % of 17 year olds passing at least five GCSE/O-levels with grade A-C, by survey year

0

10

20

30

40

50

60

1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996

Boys

Girls

Conclusion• Descriptive analysis shows girls doing better over

time, and that the change in gender gaps largely driven by changes at secondary school level.

• Shift in gender differential after the introduction of GCSEs – likely to be the most important explanation for growth in the gender gap and suggests policy can matter.

• Testing this further? - Most straightforward way would be to compare English students to students of another comparable exam system (Republic of Ireland?) before and after the introduction of GCSEs in England