Labour market expectations, relative performance and subject choice

Nuffield
Timeframe
2011-2013

Principal Investigator
Professor Peter Davies
p.davies.1@bham.ac.uk

Since the quality of educational outcomes is crucial for future prosperity and well-being the subject choices made by students are ‘high stakes’ for others as well as themselves. This project, funded by the Nuffield Foundation, evaluates the effects of providing 15-16 year-old students with information about the differences between earnings of graduates from different subjects.

Early specialisation remains a distinctive attribute of the English education system, but concern is frequently expressed that specialisation within formal schooling produces too few graduates in certain subjects: notably in Science, Technology, Engineering and Mathematics (STEM) subjects and modern foreign languages. A recent report for the Nuffield Foundation (Hodgen et al. 2010) has highlighted comparatively low rates of post-16 participation in mathematics education in the UK.

Information and guidance currently available to students encourages them to choose the subjects they enjoy and the subjects they think they are good at. This advice is uncontroversial and students do choose subjects in which they have a relative advantage. However, a system which requires early specialisation needs students to have accurate expectations about the labour market implications of their choices.

This research focused on 15/16 year old students choosing subjects to study in the sixth form (Year 12). It used an intervention through which students were provided with information about graduate earnings. The effect of this intervention was then evaluated using a randomised controlled trial and the effect of the intervention measured through students’ preferences towards subjects before they start Year 12 and the actual courses they are studying in Year 12. 

Baseline Questionnaire  StudentQuestionnaire1
Intervention
Resources and Guidance for Teachers
Lesson_K_Teacher’s_notes
Control Group Resources and Guidance for Teachers Lesson_J_Teacher’s_Version

etables for 'Information and choice of A-level subjects: A cluster randomised controlled trials with linked administrative data (published in the British Educational Research Journal, 2017)

eTable 1 Descriptive statistics of the complete case sample

  N Percent Mean Standard deviation
Male 5,056 46.5%    
Expected        
Maths GCSE grade 4,406   6.59 1.22
English GCSE grade 4,390   6.51 1.06
State school 5,593 67.3%    
Ethnic group 4,504      
White   75.1%    
Other ethnic group   8.8%    
Indian   7.4%    
Pakistani   2.0%    
Black African   1.9%    
Black Caribbean   1.8%    
Chinese   1.7%    
Black Other   0.6%    
Bangladeshi   0.6%    
Father professional or managerial 4,126 59.0%    
Mother professional or managerial 4,108 44.0%    
Student aspires for professional or managerial job 4,111 83.5%    
Graduate father 3,839 51.2%    
Graduate mother 3,916 47.4%    
Eligible for free school meals 4,561 4.8%    
Salary very important for choice of subject 4,367 52.8%    
Intending to attend university        
Definitely 4,443 59.3%    
Probably 4,443 26.0%    
Unlikely 4,443 11.0%    
Definitely not 4,443 3.8%

 

 

 

eTable 2 Effect of intervention on student’s actual choice of A-levels complete case sample (N=3,594)

 

Basic adjusted

 

Fully adjusted

    95% Confidence       95% Confidence  
  Odds- interval p-   Odds- interval p-
  ratio lower upper value N ratio lower upper value
Biology 0.94 0.62 1.41 0.75 1604 0.56 0.33 0.97 0.04
Business 0.85 0.56 1.29 0.45 1603 1.25 0.84 1.85 0.27
Chemistry 1.55 1.08 2.24 0.02 1601 0.96 0.65 1.42 0.84
Computing 0.61 0.31 1.22 0.16 1596 0.46 0.26 0.80 0.01
Economics 1.26 0.83 1.92 0.28 1603 2.01 0.66 6.12 0.22
English 0.97 0.77 1.23 0.81 1602 1.21 0.81 1.81 0.35
Geography 0.98 0.70 1.37 0.92 1594 0.75 0.44 1.27 0.28
History 1.14 0.87 1.50 0.34 1594 1.54 0.94 2.52 0.09
Languages 1.18 0.84 1.66 0.34 1604 1.15 0.69 1.94 0.59
Maths 1.58 1.08 2.31 0.02 1611 1.52 0.96 2.41 0.08
Physics 1.48 1.10 1.98 0.01 1598 1.24 0.81 1.91 0.32
Psychology 0.76 0.48 1.20 0.23 1602 0.95 0.65 1.37 0.77

Note: standard errors clustered by school. Basic adjusted result adjust for stratification variables: private school, above and below median achieving school and single sex school. Fully adjusted results additional adjust for KS3 achievement in English, maths, and science, expected maths and English GCSE grades (A*=8, G=1), gender and free school meal status.

eTable 3 Students’ Earnings Expectations Prior to Intervention (£k) Complete Case

 

Average

Lower limit

Higher limit

Me

  N Mean N Mean N Mean N Mean
Art 3,930 27.02 3,825 17.40 3,814 42.47 3,086 22.47
Business 3,875 36.24 3,763 25.11 3,770 50.35 3,086 35.27
Education 3,840 30.72 3,732 21.16 3,720 43.10 3,046 30.60
Engineering 3,822 35.77 3,726 24.99 3,716 48.18 3,044 33.81
History 3,797 31.68 3,713 21.82 3,690 42.67 3,010 29.00
Languages 3,808 32.59 3,712 22.42 3,694 43.86 2,997 29.01
Law 3,813 45.02 3,726 32.44 3,762 58.51 3,037 45.01
Maths 3,838 38.56 3,730 26.88 3,741 51.40 2,993 37.71
Medicine 3,813 44.51 3,734 31.98 3,745 58.04 3,015 44.09
Physics 3,808 38.93 3,728 27.60 3,732 51.28 2,974 36.79
Politics 3,792 38.63 3,709 27.12 3,712 51.49 3,005 36.54

eTable 5 Change in Expectations of Graduate Salaries (£k, complete case sample)

 

Average salaries

My salary
     

95% Confidence

     

95% Confidence

 
    Mean

interval

P-   Mean

interval

P-
  N difference Lower Upper value N difference CI CI value
Art 1,845 -0.77 -1.77 0.24 0.13 1222 1.72 0.52 2.93 0.006
Business 1,834 -0.83 -2.19 0.54 0.23 1260 0.50 -1.28 2.27 0.58
Education 1,829 1.22 0.37 2.07 0.006 1257 2.80 1.50 4.11 9.6E-5
Engineering 1,830 0.42 -0.75 1.60 0.47 1255 2.39 0.62 4.17 0.010
History 1,828 -1.82 -3.03 -0.60 0.004 1249 0.58 -0.67 1.82 0.35
Languages 1,822 -2.46 -3.84 -1.07 9.1E-4 1239 -1.68 -3.72 0.35 0.10
Law 1,827 -3.50 -4.85 -2.15 5.8E-6 1256 -1.92 -3.84 0.00 0.05
Maths 1,831 -0.32 -1.71 1.07 0.65 1242 2.18 0.45 3.92 0.02
Medicine 1,828 -2.70 -3.86 -1.53 3.3E-5 1243 -1.97 -3.63 -0.31 0.02
Physics 1,825 -3.04 -4.50 -1.58 1.4E-4 1246 -0.54 -2.08 1.01 0.49
Politics 1,808 -3.48 -4.80 -2.15 4.7E-6 1246 -2.15 -4.09 -0.20 0.03

Note: standard errors clustered by school. Basic adjusted result adjust for stratification variables: private school, above and below median achieving school and single sex school. Fully adjusted results additional adjust for KS3 achievement in English, maths, and science, expected maths and English GCSE grades (A*=8, G=1), gender and free school meal status.

eTable 6 Effect of intervention on students’ intentions to study Biology, Chemistry, Computing, Math and Physics, complete case sample

    Basic adjusted   Fully adjusted
      95% Confidence       95% Confidence  
    Mean interval P-   Mean interval P-
  N difference Lower Upper value N difference Lower Upper value
Biology 2,234 -0.13 -0.27 0.00 0.04 1,158 -0.13 -0.42 0.16 0.37
Business 2,224 0.02 -0.09 0.14 0.66 1,152 -0.02 -0.16 0.13 0.82
Chemistry 2,230 -0.06 -0.18 0.05 0.26 1,160 -0.15 -0.28 -0.02 0.03
Computing 2,200 0.12 0.03 0.22 0.01 1,149 -0.06 -0.17 0.04 0.22
Economics 2,207 0.00 -0.10 0.10 0.97 1,143 0.04 -0.10 0.18 0.56
English 2,225 -0.15 -0.27 -0.04 0.01 1,156 -0.18 -0.36 0.00 0.05
Geography 2,210 -0.10 -0.18 -0.01 0.03 1,143 -0.16 -0.27 -0.06 0.004
History 2,207 -0.01 -0.10 0.08 0.82 1,148 -0.08 -0.23 0.07 0.27
Languages 2,229 -0.17 -0.29 -0.06 0.004 1,157 -0.04 -0.21 0.13 0.59
Maths 2,241 0.03 -0.08 0.15 0.58 1,166 -0.08 -0.28 0.12 0.40
Physics 2,207 0.02 -0.07 0.11 0.61 1,143 -0.14 -0.26 -0.03 0.02
Psychology 2,205 0.03 -0.13 0.19 0.69 1,142 0.05 -0.20 0.31 0.67

Note: standard errors clustered by school. Basic adjusted result adjust for stratification variables: private school, above and below median achieving school and single sex school. Fully adjusted results additional adjust for KS3 achievement in English, maths, and science, expected maths and English GCSE grades (A*=8, G=1), gender and free school meal status.

eTable 7 P-values of the effect of the intervention on actual subject choice allowing for bivariate correlations between subjects (N=5,593)

Allowing P-value on effect of                
joint Biolog Busin Chemi Compu Econo Engl Geogra Hist Langu Mat Phys Psycho
Biology   0.86 0.39 0.03 0.50 0.48 0.85 0.17 0.52 0.04 0.79 0.39
Business 0.05   0.41 0.03 0.47 0.46 0.85 0.15 0.51 0.04 0.76 0.40
Chemistry 0.06 0.92   0.03 0.53 0.53 0.85 0.18 0.54 0.03 0.71 0.37
Computing 0.06 0.86 0.45   0.51 0.45 0.84 0.16 0.54 0.04 0.80 0.39
Economics 0.05 0.86 0.46 0.03   0.45 0.86 0.16 0.52 0.04 0.78 0.39
English 0.05 0.87 0.49 0.03 0.50   0.85 0.14 0.51 0.04 0.82 0.39
Geography 0.05 0.85 0.43 0.03 0.49 0.47   0.16 0.51 0.04 0.77 0.39
History 0.05 0.85 0.43 0.03 0.49 0.45 0.84   0.52 0.04 0.78 0.39
Languages 0.05 0.85 0.43 0.03 0.49 0.46 0.85 0.16   0.04 0.78 0.40
Maths 0.05 0.87 0.41 0.03 0.49 0.45 0.82 0.17 0.53   0.76 0.40
Physics 0.05 0.88 0.45 0.03 0.49 0.48 0.84 0.17 0.52 0.04   0.38
Psychology 0.05 0.86 0.44 0.03 0.49 0.46 0.86 0.16 0.52 0.04 0.81  

Note: Estimated using mvprobit in Stata and the imputed dataset. Each column is the p-value from the coefficient on the intervention allowing for the joint choice of the subjects in the left column. The p values for Maths all below .001 and truncated because the column space is narrow. The p-values are not affected by relaxing the assumption of independence of choice of subjects. Therefore the study’s conclusions are dependent on assuming independence of choice of subject. Standard errors clustered by school. Basic adjusted results adjust for stratification variables: private school, above and below median achieving school and single sex school. Fully adjusted results additional adjust for KS3 achievement in English, maths, and science, expected maths and English GCSE grades (A*=8, G=1), gender and free school meal status. 

eTable 8 Effect of intervention on A-level choice adjusted for Key Stage 2 results (N=5,593)

 

Fully adjusted for Key Stage 2 results

    95% Confidence  
    intervals  
A-level chosen Odds-ratio Lower Upper P value
Biology 0.80 0.60 1.06 0.11
Business 0.98 0.72 1.33 0.90
Chemistry 1.22 0.93 1.59 0.15
Computing 0.59 0.38 0.92 0.02
Economics 1.13 0.80 1.59 0.48
English 1.11 0.89 1.39 0.36
Geography 0.99 0.73 1.36 0.97
History 1.23 0.96 1.59 0.11
Languages 1.13 0.82 1.57 0.46
Maths 1.44 1.09 1.90 0.01
Physics 1.12 0.85 1.47 0.41
Psychology 0.89 0.64 1.23 0.48

Note: Uses imputed dataset. Standard errors clustered by school. Basic adjusted result adjust for stratification variables: private school, above and below median achieving school and single sex school. Fully adjusted results additional adjust for KS2 achievement in English, maths, and science, expected maths and English GCSE grades (A*=8, G=1), gender and free school meal status.

eTable 9 Heterogeneity in the effect of the intervention between state and private schools(n= 5,593)

 

 

Basic adjusted

Fully adjusted

     

95% Confidence

   

95% Confidence

 
    Odds-

intervals

P- Odds-

intervals

P-
    ratio Lower Upper value ratio Lower Upper value
Biology Main effect 0.93 0.55 1.56 0.77 0.66 0.44 1.00 0.05
  Private school interaction 0.95 0.49 1.82 0.87 1.24 0.77 1.99 0.38
Business Main effect 0.87 0.62 1.21 0.40 0.94 0.69 1.29 0.71
  Private school interaction 1.38 0.53 3.60 0.51 1.28 0.58 2.85 0.54
Chemistry Main effect 1.36 0.83 2.23 0.22 0.97 0.70 1.36 0.87
  Private school interaction 0.98 0.57 1.70 0.95 1.24 0.79 1.95 0.34
Computing Main effect 0.65 0.39 1.08 0.10 0.60 0.38 0.94 0.03
  Private school interaction 0.64 0.16 2.56 0.53 0.57 0.15 2.17 0.41
Economics Main effect 1.91 0.90 4.05

0.09

1.42

0.73

2.74

0.30

 

Private school interaction

0.54

0.23

1.26

0.15

0.67

0.32

1.39

0.28

English

Main effect

0.96

0.74

1.24

0.75

1.08

0.81

1.45

0.61

 

Private school interaction

0.91

0.63

1.33

0.64

1.00

0.67

1.51

0.99

Geography Main effect 0.95 0.61 1.48 0.83 0.78 0.51 1.20 0.26
  Private school interaction 1.08 0.60 1.92 0.80 1.48 0.84 2.60 0.17

History

Main effect

1.27

0.88

1.83

0.20

1.23

0.87

1.75

0.24

 

Private school interaction

0.83

0.49

1.39

0.47

0.99

0.60

1.64

0.97

Languages

Main effect

1.27

0.71

2.27

0.43

1.11

0.69

1.80

0.67

 

Private school interaction

0.82

0.42

1.60

0.56

1.00

0.55

1.81

1.00

Maths

Main effect

1.67

1.06

2.63

0.03

1.38

0.97

1.98

0.08

 

Private school interaction

0.71

0.38

1.32

0.28

0.86

0.52

1.44

0.57

Physics

Main effect

1.49

0.95

2.35

0.09

1.03

0.69

1.54

0.89

 

Private school interaction

0.82

0.45

1.47

0.50

0.96

0.57

1.60

0.86

Psychology

Main effect

0.75

0.54

1.05

0.10

0.85

0.63

1.14

0.27

 

Private school interaction

1.13

0.62

2.08

0.69

1.03

0.50

2.14

0.93

Note: Uses imputed dataset. Standard errors clustered by school. Basic adjusted result adjust for stratification variables: private school, above and below median achieving school and single sex school. Fully adjusted results additional adjust for KS3 achievement in English, maths, and science, expected maths and English GCSE grades (A*=8, G=1), gender and free school meal status, and the interaction of each of these covariates with the private school indicator.

eTable 10 Heterogeneity in the effect of the intervention between male and female students schools (N=5,593).

    Basic adjusted >Fully adjusted
      95% Confidence     95% Confidence  
    Odds- interval P- Odds- interval P-
    ratio Lower Upper value ratio Lower Upper value
Biology Overall effect 0.95 0.65 1.40 0.80 0.71 0.51 0.99 0.05
  Male interaction 1.04 0.70 1.53 0.86 1.18 0.80 1.73 0.40
Business Overall effect 1.03 0.68 1.56 0.89 1.13 0.76 1.68 0.53
  Male interaction 0.77 0.48 1.23 0.27 0.80 0.49 1.30 0.37
Chemistry Overall effect 1.29 0.86 1.93 0.21 1.03 0.75 1.42 0.86
  Male interaction 1.21 0.82 1.79 0.34 1.24 0.85 1.80 0.26
Computing Overall effect 0.63 0.35 1.13 0.12 0.64 0.36 1.16 0.14
  Male interaction 0.84 0.43 1.66 0.62 0.87 0.44 1.73 0.70
Economics Overall effect 1.15 0.71 1.86 0.57 1.16 0.73 1.86 0.52
  Male interaction 1.09 0.65 1.83 0.75 0.93 0.55 1.57 0.77
English Overall effect 1.18 0.94 1.47 0.15 1.21 0.93 1.58 0.16
  Male interaction 0.69 0.52 0.91 0.010 0.76 0.53 1.09 0.14
Geography Overall effect 0.91 0.62 1.33 0.63 0.81 0.54 1.22 0.32
  Male interaction 1.13 0.69 1.85 0.64 1.44 0.85 2.42 0.17
History Overall effect 1.11 0.79 1.55 0.55 1.22 0.86 1.74 0.27
  Male interaction 1.08 0.76 1.54 0.65 1.05 0.72 1.52 0.80
Languages Overall effect 1.19 0.81 1.75 0.37 1.05 0.74 1.49 0.79
  Male interaction 0.91 0.55 1.49 0.70 1.08 0.65 1.77 0.77
Maths Overall effect 1.34 0.95 1.89 0.09 1.38 1.00 1.90 0.05
  Male interaction 1.08 0.73 1.59 0.69 0.88 0.61 1.29 0.52
Physics Overall effect 1.26 0.78 2.04 0.35 1.10 0.74 1.65 0.63
  Male interaction 1.03 0.60 1.76 0.91 0.96 0.56 1.64 0.88
Psychology Overall effect 0.85 0.61 1.18 0.34 0.81 0.58 1.14 0.22
  Male interaction 1.17 0.66 2.11 0.59 1.23 0.68 2.23 0.50

Note: Uses imputed dataset. Standard errors clustered by school. Basic adjusted result adjust for stratification variables: private school, above and below median achieving school and single sex school. Fully adjusted results additional adjust for KS3 achievement in English, maths, and science, expected maths and English GCSE grades (A*=8, G=1), gender and free school meal status, and the interaction of each of these covariates with the private school indicator.

eTable 11 Intraclass correlation of subject choice (N=3,594).

A-level chosen Intraclass correlation
Biology 0.089
Business 0.119
Chemistry 0.097
Computing 0.240
Economics 0.113
English 0.043
Geography 0.045
History 0.048
Languages 0.067
Maths 0.125
Physics 0.056
Psychology 0.118

Publications

Davies, P. and Qiu, T. (2016) Labour market expectations, relative performance and subject choice, Research report, University of Birmingham.

Davies, P., Qiu, T. and Davies, N. (2014). Cultural and Human Capital, Information and Higher Education Choices, Journal of Education Policy ifirst. DOI:10.1080/02680939.2014.891762