Intergenerational educational mobility – The role of non-cognitive skills

Figures & data

Table 1. Next steps data used to construct non-cognitive skill measures.

	How many times was the question asked in Waves 1–4 from the same individual?	Wave 1 (age 13)	Wave 2 (age 14)	Wave 3 (age 15)	Wave 4 (age 16)	Missing or no answer
Locus of control
Q1	Once		6146			1390
Q2	Once		6698			838
Q3	Once		6213			1323
Q4	Once		6260			1276
Q5	Once		6392			1144
Q6	Once		5953			1583
Academic self-concept
Q1	Three times^a	7115	286	53		82
Q2	Once	7308				228
Q3	Once	7434				102
Q4	Once	7432				104
Q5	Once	7427				109
Work ethic
Q1	Three times^b	7177	271	32		56
Q2	Once		6667			869
Q3	Once		6696			840
Self-esteem
Q1	Two times		6133		6246	703
Q2	Two times		6641		6306	607

Source: University College London, UCL Institute of Education, Centre for Longitudinal Studies. (2018).

^a Q1 for academic self-concept was asked three times. From Waves 2 and 3, we only use data for individuals with missing data in Wave 1 (although data are available for more people).

^b Q1 for work ethic was asked three times. From Waves 2 and 3, we only use data for individuals with missing data in Wave 1 (although data are available for more people). Sample of those present at Wave 8.

Table 2. Sample descriptive statistics: means and proportions of variables of interest.

	Total sample			Complete case sample			Difference between the total and the complete case sample	t-test p-value
	Obs	Mean	SD	Obs	Mean	SD
All individuals
External locus of control (continuous, standardized)	4937	0	1	4755	−0.02	0.99	0.02	0.26
Academic self-concept (continuous, standardized)	7259	0	1	4755	0.06	0.98	−0.06***	0.00
Work ethic (continuous, standardized)	6538	0	1	4755	0.02	0.99	−0.02	0.39
Low self-esteem (binary)	7507	0.11	0.31	4755	0.11	0.31	0.00	0.76
HE participation (binary)	7536	0.40	0.49	4755	0.42	0.49	−0.02***	0.01
Potential First-generation: young people whose parents do not have university degrees
External locus of control (continuous, standardized)	3906	0.05	1.01	3753	0.03	1.00	0.02	0.29
Academic self-concept (continuous, standardized)	5834	−0.06	0.99	3753	0.00	0.96	−0.06***	0.00
Work ethic (continuous, standardized)	5211	−0.01	1.00	3753	0.00	0.99	−0.02	0.46
Low self-esteem (binary)	6048	0.11	0.32	3753	0.12	0.32	0.00	0.52
HE participation (binary)	6073	0.33	0.47	3753	0.35	0.48	−0.02**	0.05
Children of graduate parents
External locus of control (continuous, standardized)	1031	−0.23	0.93	1002	−0.24	0.93	0.01	0.77
Academic self-concept (continuous, standardized)	1425	0.31	1.01	1002	0.37	1.00	−0.06	0.17
Work ethic (continuous, standardized)	1327	0.06	0.98	1002	0.08	0.98	−0.02	0.69
Low self-esteem (binary)	1459	0.09	0.29	1002	0.08	0.28	−0.01	0.52
HE participation (binary)	1463	0.72	0.45	1002	0.73	0.44	−0.01	0.51

Notes: ***p < 0.01, **p < 0.05, *p < 0.1 Weighted using Wave 8 weights.Source: University College London, UCL Institute of Education, Centre for Longitudinal Studies. (2018).

Table 3. Comparison groups used in this paper.

	Potential first-generation individuals (children of non-graduate parents)	Children of graduate parents
Young people who did not go to university	Group 1: Those matching their parents with no HE (N = 2581)	Group 2: Downwardly mobile group (N = 450)
Young people who went to university	Group 3: First-generation students (N = 1172)	Group 4: Those matching their parents with HE (N = 552)

The number of observations (N) refers to the complete case sample.

Source: University College London, UCL Institute of Education, Centre for Longitudinal Studies. (2018).

Figure 1. Non-cognitive skills, raw and conditional means.

Notes: N = 4937, 6538, 7259 and 7507, respectively. Weighted using Wave 8 weights. Conditional means control for gender, age, number of siblings, region, whether the young person is a care leaver, whether the young person had a low birth weight or was born prematurely, ethnicity, mother’s social class, father’s social class, age of mother, age of father, single parent household, income deprivation, disability, whether their grandparent attended university, Free School Meal (FSM) status, private school, the highest qualification of the mother and father, and Key Stage 2 quintile of achievement. Source: University College London, UCL Institute of Education, Centre for Longitudinal Studies. (2018).

Table 4. The relationship between non-cognitive skills and university attendance among potential first-generation individuals.

	(1)	(2)	(3)	(4)	(5)	(6)
External locus of control (continuous, standardized)	−0.071***				−0.038***	−0.003
	(0.008)				(0.008)	(0.008)
Work ethic (continuous, standardized)		0.071***			0.025***	0.017**
		(0.008)			(0.008)	(0.008)
Academic self-concept (continuous, standardized)			0.110***		0.089***	0.028***
			(0.009)		(0.009)	(0.008)
Low self-esteem (binary)				−0.099***	−0.054**	−0.033
				(0.023)	(0.023)	(0.022)
Constant	1.308**	1.389**	1.246**	1.436**	1.208**	1.152**
	(0.556)	(0.566)	(0.556)	(0.572)	(0.539)	(0.501)
Observations	3753	3753	3753	3753	3753	3753

Notes: Sample of potential first in family individuals (i.e. neither parent has a university degree). Robust standard errors clustered by school in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. Weighted using Wave 8 weights. All models control for gender, age, number of siblings, region, whether the young person is a care leaver, whether the young person had a low birth weight or was born prematurely, ethnicity, mother’s social class, father’s social class, age of mother, age of father, single parent household, income deprivation, disability, whether their grandparent attended university, Free School Meal (FSM) status, private school, the highest qualification of the mother and father, and Key Stage 2 quintile of achievement. Column 6 additionally includes GCSE total points quintile. Model 5 using CFA instead of PCA to create the indexes of non-cognitive skills is reported in Table O1 in the Online Appendix. Source: University College London, UCL Institute of Education, Centre for Longitudinal Studies. (2018).

Figure 2. The gender gap in non-cognitive skills.

Notes: Positive gaps indicate that women have higher level of a skill. Total N = 4937, 6538, 7259 and 7507, respectively. Weighted using Wave 8 weights. Conditional gender wage gap models control for age, number of siblings, region, whether the young person is a care leaver, whether the young person had a low birth weight or was born prematurely, ethnicity, mother’s social class, father’s social class, age of mother, age of father, single parent household, income deprivation, disability, whether their grandparent attended university, Free School Meal (FSM) status, private school, the highest school qualification of the mother and father, and their Key Stage 2 quintile of achievement. Source: University College London, UCL Institute of Education, Centre for Longitudinal Studies. (2018).

Table 5. The relationship between non-cognitive skills and university attendance, by gender*non-cognitive skills interactions (sample of potential first-generation individuals).

	(1)	(2)	(3)	(4)	(5)	(6)
Female	0.076***	0.064***	0.086***	0.083***	0.087***	0.036**
	(0.016)	(0.016)	(0.016)	(0.018)	(0.017)	(0.016)
External locus of control (continuous, standardized)	−0.055***				−0.022*	0.010
	(0.011)				(0.012)	(0.012)
Female* External locus of control	−0.035**				−0.037**	−0.028*
	(0.016)				(0.017)	(0.016)
Work ethic (continuous, standardized)		0.072***			0.022*	0.013
		(0.012)			(0.013)	(0.012)
Female* Work ethic		−0.003			0.003	0.008
		(0.015)			(0.017)	(0.015)
Academic self-concept (continuous, standardized)			0.134***		0.121***	0.051***
			(0.012)		(0.014)	(0.012)
Female*Academic self-concept			−0.046***		−0.062***	−0.043***
			(0.016)		(0.018)	(0.015)
Low self-esteem (binary)				−0.089**	−0.034	−0.014
				(0.037)	(0.037)	(0.034)
Female* Low self-esteem				−0.016	−0.026	−0.026
				(0.044)	(0.044)	(0.041)
Constant	1.301**	1.387**	1.259**	1.435**	1.217**	1.159**
	(0.556)	(0.566)	(0.557)	(0.572)	(0.540)	(0.502)
Observations	3,753	3,753	3,753	3,753	3,753	3,753

Notes: Sample of potential first in family individuals (i.e. neither parent has a university degree). Robust standard errors clustered by school in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. Weighted using Wave 8 weights. All models control for age, number of siblings, region, whether the young person is a care leaver, whether the young person had a low birth weight or was born prematurely, ethnicity, mother’s social class, father’s social class, age of mother, age of father, single parent household, income deprivation, disability, whether their grandparent attended university, Free School Meal (FSM) status, private school, the highest qualification of the mother and father, and Key Stage 2 quintile of achievement. Column 6 additionally includes GCSE total points quintile. Model 5 using CFA instead of PCA to create the indexes of non-cognitive skills is reported in Table O2 in the Online Appendix. Source: University College London, UCL Institute of Education, Centre for Longitudinal Studies. (2018).

Table 6. Kitagawa-Blinder-Oaxaca decomposition of university attendance among potential first-generation individuals and children of graduate parents.

	Total sample	Boys	Girls
Group 1: children of graduate parents	0.773***	0.758***	0.789***
	(0.013)	(0.019)	(0.018)
Group 2: potential first-generation individuals	0.473***	0.430***	0.507***
	(0.008)	(0.012)	(0.011)
Difference between the two groups	0.300***	0.327***	0.282***
	(0.016)	(0.023)	(0.021)
Explained part	0.187***	0.194***	0.178***
	(0.012)	(0.019)	(0.017)
Unexplained part	0.114***	0.134***	0.105***
	(0.018)	(0.026)	(0.024)
Explained difference
External locus of control	0.009***	0.005	0.010***
(continuous, standardized)	(0.002)	(0.003)	(0.003)
Work ethic	0.001	0.002	0.001
(continuous, standardized)	(0.001)	(0.001)	(0.001)
Academic self-concept	0.031***	0.043***	0.020***
(continuous, standardized)	(0.004)	(0.007)	(0.004)
Low self-esteem	0.001	0.001	0.001
(binary)	(0.001)	(0.001)	(0.001)
KS2	0.020***	0.019***	0.021***
	(0.005)	(0.006)	(0.007)
Family	0.024***	0.020***	0.027***
	(0.005)	(0.008)	(0.007)
Region	0.001	0.004	−0.002
	(0.002)	(0.003)	(0.002)
Disadvantage	0.016***	0.014**	0.015**
	(0.005)	(0.006)	(0.007)
Birth	−0.019***	−0.015**	−0.017***
	(0.005)	(0.006)	(0.006)
Parents	0.082***	0.087***	0.075***
	(0.011)	(0.017)	(0.015)
Independent school	0.020***	0.014**	0.027***
	(0.005)	(0.006)	(0.007)
Observations	4755	2150	2605

Notes: Robust standard errors clustered by school in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. Weighted using Wave 8 weights. All models control for the same variables as before, grouped into the following categories: Region; Birth: a set of dummies with birth weight category, a dummy variable for a premature birth, ethnicity dummies, and the age of the mother at birth; Family: the number of siblings, single parent household dummy, and a dummy for whether or not the individual’s grandparents had a university degree; Disadvantage: a dummy for FSM eligibility, a dummy for being a care leaver, a dummy for having a disability, and dummies for household income deprivation; Parents: mother’s and father’s social class and age; Key Stage 2: quintiles of age 11 test scores. ‘Group 1’ refers to individuals whose parents have a university degree and ‘Group 2’ refers to those individuals whose parents do not have a university degree. Source: University College London, UCL Institute of Education, Centre for Longitudinal Studies. (2018).

Table 7. Kitagawa-Blinder-Oaxaca decomposition of university attendance among potential first-generation individuals and children of graduate parents: the role of age 16 test (GCSE) scores.

	University participation			GCSE’s (standardized)
	Total	Boys	Girls	Total	Boys	Girls
Group 1: children of graduate parents	0.773***	0.758***	0.789***	0.795***	0.754***	0.835***
	(0.013)	(0.019)	(0.018)	(0.019)	(0.026)	(0.028)
Group 2: potential first-generation individuals	0.473***	0.430***	0.507***	0.243***	0.142***	0.324***
	(0.008)	(0.012)	(0.011)	(0.013)	(0.021)	(0.017)
Difference between the two groups	0.300***	0.327***	0.282***	0.551***	0.613***	0.511***
	(0.016)	(0.023)	(0.021)	(0.023)	(0.033)	(0.033)
Explained part	0.239***	0.255***	0.224***	0.366***	0.398***	0.341***
	(0.013)	(0.019)	(0.018)	(0.021)	(0.034)	(0.028)
Unexplained part	0.062***	0.072***	0.058***	0.186***	0.214***	0.170***
	(0.017)	(0.025)	(0.022)	(0.022)	(0.033)	(0.031)
Explained difference
External locus of control	0.000	−0.004	0.002	0.032***	0.042***	0.023***
(continuous, standardized)	(0.002)	(0.003)	(0.002)	(0.006)	(0.009)	(0.007)
Work ethic	0.001	0.001	0.001	0.001	0.002	0.000
(continuous, standardized)	(0.001)	(0.001)	(0.001)	(0.001)	(0.002)	(0.001)
Academic self-concept	0.009***	0.017***	0.002	0.083***	0.108***	0.061***
(continuous, standardized)	(0.003)	(0.005)	(0.003)	(0.009)	(0.015)	(0.011)
Low self-esteem	0.001	0.000	0.001	0.001	0.001	0.001
(binary)	(0.000)	(0.001)	(0.001)	(0.001)	(0.001)	(0.002)
KS2	0.007	0.006	0.006	0.043***	0.045***	0.042***
	(0.004)	(0.006)	(0.006)	(0.006)	(0.009)	(0.009)
GCSE	0.148***	0.161***	0.139***
	(0.008)	(0.013)	(0.011)
Family	0.016***	0.014*	0.016**	0.039***	0.029***	0.046***
	(0.005)	(0.007)	(0.007)	(0.007)	(0.011)	(0.010)
Region	−0.001	0.002	−0.002	0.004	0.007	0.001
	(0.001)	(0.003)	(0.002)	(0.002)	(0.005)	(0.003)
Disadvantage	0.002	0.001	−0.001	0.013	0.013	0.011
	(0.005)	(0.007)	(0.008)	(0.010)	(0.014)	(0.014)
Birth	−0.013***	−0.014**	−0.011**	−0.023***	−0.009	−0.016**
	(0.004)	(0.006)	(0.005)	(0.006)	(0.006)	(0.006)
Parents	0.039***	0.044***	0.034**	0.149***	0.148***	0.140***
	(0.010)	(0.016)	(0.014)	(0.015)	(0.024)	(0.021)
Independent school	0.031***	0.026***	0.037***	0.023***	0.012	0.032**
	(0.006)	(0.008)	(0.008)	(0.008)	(0.010)	(0.013)
Observations	4,755	2,150	2,605	4,420	2,005	2,415

Notes: Robust standard errors clustered by school in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. Weighted using Wave 8 weights. All models control for the same variables as before, grouped into the following categories: Region; Birth: a set of dummies with birth weight category, a dummy variable for a premature birth, ethnicity dummies, and the age of the mother at birth; Family: the number of siblings, single parent household dummy, and a dummy for whether or not the individual’s grandparents had a university degree; Disadvantage: a dummy for FSM eligibility, a dummy for being a care leaver, a dummy for having a disability, and dummies for household income deprivation; Parents: mother’s and father’s social class and age; Key Stage 2: quintiles of age 11 test scores. ‘Group 1’ refers to individuals whose parents have a university degree and ‘Group 2’ refers to those individuals whose parents do not have a university degree. Source: University College London, UCL Institute of Education, Centre for Longitudinal Studies. (2018).

University College London, UCL Institute of Education, Centre for Longitudinal Studies. 2018. Next Steps: Sweeps 1-8, 2004-2016: Secure Access [data collection]. doi:10.5255/UKDA-SN-7104-4.

 Google Scholar