Table 1.

Summary of variables.

Longitudinal Study of Australian Children (LSAC)		Millennium Cohort Study (MCS)
Socio-economic disadvantage (SED): Maternal education (0–1 years)		Socio-economic disadvantage (SED): Maternal education (9 months)
Mothers' highest level of educational attainment	•Low: Did not complete year 12a •High: Completed year 12, certificate/diploma, degree	Mothers' highest academic qualification	•Low: GCSEb grades D-G, or below •High: GCSE grades A*-C, A-Levelsc, Diploma, Degree
Academic achievement (8–9 years)		Academic achievement (7 years)
Maths	•Maths domain of Academic Rating Scale (ARS), teacher report (Australian Institute of Family Studies, 2011, Rothman, 2009) •Age standardised scores were divided into quintiles, with the bottom quintile representing ‘low achievement’	Maths	•Shortened version of the National Foundation for Education Research standard Progress in Maths test (Connelly, 2013) •Completed by the cohort child •Scores were divided into quintiles, with the bottom quintile representing ‘low achievement’ •Quintiles were assigned within each school year because scores were not age standardised (Connelly, 2013) (1% children were in ‘year 1’, 94% ‘year 2’, 5% ‘year 3’)
Literacy	•Literacy domain of the ARS, completed by the teacher (Australian Institute of Family Studies, 2011, Rothman, 2009). •Scores were standardised in a Rasch model, and divided into quintiles •‘Low achievement’ = bottom quintile of scores	Literacy	•British Ability Scales II (BAS II) subtests for word reading, completed by the cohort child (Connelly, 2013) •Age standardised scores were divided into quintiles •‘Low achievement’ = bottom quintile of scores
Self-regulation (6–7 years)		Self-regulation (5 years)
Task attentiveness and persistence	•Previously created measure Sawyer et al., 2015 consisting of: •Five items from the Short Temperament Scale (When child starts a project… he/she works on it without stopping until it is completed…; likes to complete one task or activity before going onto the next; stays with an activity for a long time; when a toy or game is difficult, quickly turns to another activity) (Fullard et al., 1984) •One item from the Strengths and Difficulties Questionnaire (Sees tasks through to the end, has good attention span) (Goodman, 2001)) •Bottom quintile of scores = ‘low self-regulation’	Task attentiveness and persistence	•Independence and Self-regulation domain of the Child Social Behaviour Questionnaire (EPPE) (Likes to work things out for self; Does not need much help with tasks; Chooses activities on own; Persists in the face of difficult tasks; Move to new activity after finishing task) (Johnson et al., 2012) •Bottom quintile of scores = ‘low self-regulation’
Cognitive ability (6–7 years)		Cognitive ability (5 years)
Non-verbal & verbal ability	•Non-verbal ability: Matrix Reasoning subtest of the Wechsler Intelligence Scale for Children, IV Edition (Wechsler, 2003) •Verbal ability: the Peabody Picture Vocabulary Test (PPVT)-III - LSAC Australian Short-form (Australian Institute of Family Studies, 2011, Rothman, 2005) •Age standardised non-verbal and verbal scores were combined and converted to T-scores, as recommended when using multiple cognitive ability scales (Connelly, 2013) •Bottom quintile of scores = low cognitive ability	Non-verbal & verbal ability	•Non-verbal ability: pattern construction subtest of the BAS II(Connelly, 2013) •Verbal ability: BAS II naming vocabulary subtest (Connelly, 2013) •Age standardised non-verbal and verbal scores were combined, converted to T-scores, as recommended when using multiple cognitive ability scales (Connelly, 2013) •Bottom quintile of scores = low cognitive ability
Baseline confounders (0–1 years)		Baseline confounders (9 months)
Maternal young age at first live birth	Age at first live birth was not directly captured in LSAC, and was estimated using the following: •Mothers' age (years) at the birth of study child •Age (in years) of all other children living in the household •Mother's relationship to these children •Age of mother's eldest biological non-resident child (0−2y, 3−4y, 5−10y, 11−17y, 18y+) •Estimated age at first live birth: <20 years, > = 20 years	Maternal young age at first live birth	•Age at first live birth (years): <20 years, > = 20 years
Language	•Main language spoken at home by the mother with the study child, coded as ‘English’ or ‘other’	Language	•Main household language, coded as ‘English’ or ‘other’
Ethnicityd	N/A	Ethnicityd	•White, Black, Indian, Pakistani/Bangladeshi, Mixed, Other
Intermediate confounding (various ages)		Intermediate confounding (various ages)
0–1 years	•Whether the mother drank alcohol during pregnancy (yes, no)	9 months	•Whether the mother drank alcohol during pregnancy (yes, no)
•Whether the mother smoked cigarettes during pregnancy (yes, no)		•Whether the mother smoked cigarettes during pregnancy (yes, no)
2–3 years	•Formal childcare use (daycare, preschool or kindergarten)
4–5 years	•Lone parent family •Housing tenure: ‘owned/mortgaged’, ‘renting or other’ •Weekly household income, divided into quintiles •Workless household (no parent in paid employment) •Maternal psychological distress (score>13, Kessler K6 (Kessler et al., 2002)) •Parenting warmth towards the child reported by mother, across six items (Australian Institute of Family Studies, 2011). Scores were highly skewed so lower warmth was defined as < median	3 years	•Formal childcare (nursery, childcare centre, or registered childminder) •Lone parent family •Housing tenure: ‘owned/mortgaged’, ‘privately renting’, ‘socially renting or other'e •Equivalised weekly household income (using a modified OECD equivalence scale (Bradshaw and Holmes, 2010)), in quintiles •Workless household (no parent in paid employment) •Maternal psychological distress (score>13, Kessler K6 (Johnson et al., 2012, Kessler et al., 2002)) •Parenting warmth (Pianta scale, ranging from 0 to 35); lower warmth was defined as < 30 (Johnson et al., 2012)

^aYear 12: indicates completion of high school.

^bGCSE: General Certificate of Secondary Education, received upon completion of high school exams (at approx. age 16 years).

^cA-Levels: General Certificate of Education Advanced Level, received upon completion of higher education (approx. 18 years).

^dEthnicity was not collected in LSAC and so was used in a sensitivity analysis of MCS data only.

^eThe MCS housing tenure variable differentiated between private and social renting, due to large differences in the socio-economic disadvantage and housing needs of these two groups in the UK. The LSAC question on housing tenure does not allow this distinction, the differences between private and social housing are less pronounced in Australia.