The Timing and Partnership Context of Becoming a Parent: Childhood Antecedents, Cohort and Gender...

Post on 28-Mar-2015

214 views 0 download

Tags:

Transcript of The Timing and Partnership Context of Becoming a Parent: Childhood Antecedents, Cohort and Gender...

The Timing and Partnership Context of Becoming a Parent: Childhood Antecedents, Cohort

and Gender

John HobcraftUniversity of York

Main Questions

• Links of Childhood Disadvantage to becoming a parent?

• How do Childhood Disadvantages play through Timing and Partnership Context?– Role in ‘off-time’ and ‘off-context’ first births

• Do these links vary by Cohort and by Gender?

Data sources

The British Birth Cohort StudiesThe National Child Development Study (NCDS)

• longitudinal study of children born in one week of March 1958: follow-ups at ages 7, 11, 16, 23, 33, and 42

British Cohort Study (BCS70)• longitudinal study of children born in one

week of April 1970: follow-ups at age 5, 10, 16, 26, and 30

• N=22,324

First Births by Age 30

Proportions having a first birth by age 30 (%)

NCDS BCS70

Males 53 39

Females 67 54

Cohort changes in exposure and birth context

Exposure to risk %

First births %

Context 1958 1970 1958 1970

Never Partnered 69 69 7 17

Out of Partnership 3 5 1 3

Cohabiting 7 16 7 30

Married ex-cohab 5 6 21 33

Married directly 16 4 64 18

Crude & standardized rates of first birth (per thousand)

Standardized for:

Crude Age Age & context

NCDS Males 47 46 46

NCDS Females 69 72 48

BCS70 Males 33 31 47

BCS70 Females 50 50 51

Changing childhood contexts

• Increasing parental divorce/ separation

• Move to earlier childbearing among CM parents

• Increased gender equity

• Residualisation of social (public) housing

Changing contexts of transition to adulthood for 1970 cf 1958

• Increased importance of education• Better average income• But greater income inequality• Rising unemployment to age 30 for 1958 cohort• Less favourable housing market• Shifts in normative partnership contexts for births• Delayed transitions to adulthood• Greater gender equity

– Narrowed education differences by gender– Increased female employment

• Lower security – employment, welfare state, partnership

Childhood Antecedents

• Socioeconomic status (4 waves)– Class of origin, Father’s class, & Child poverty

• Housing tenure (3 waves)• Parental ages at cohort member’s birth• Family disruption (4 waves)

– born out of wedlock, ever in care, parental separation, & widow(er)hood

• Parental Interest in Schooling (age 10 or 11)• Anti-Social Behaviour (3,2, & 1 waves)

– Aggression, truancy, & contact with police• Academic test scores (3 waves)

Timing and Contexts

• Age-Groups– 16-19– 20-22– 23-24– 25-29

• Partnership Contexts– Never partnered– Out of partnership– Cohabiting– Married ex-cohabiting– Direct marriage

Analytic strategy 1

• Episode files & Poisson rate models

• Basic timing model• Timing model with childhood antecedents

• Basic timing and partnership context model• Timing and partnership with childhood

antecedents

Analytic strategy 2

• Interact timing and context measures with gender & cohort and with each other

• Interact childhood antecedents with timing and with context

• Further interact all of these with gender and with cohort

• Total of 301 dummies for timing model• And 617 dummies for timing and context model

Analytic strategy 3

• Repeated ‘blocked’ stepwise regressions– Get preliminary model– Test for addition of all possible terms– Modify model to include residual extra

significant terms– Repeat until stable

• P<0.001

Results – Summary (IRRs)Timing & Context Model

• Six ‘pervasive’ childhood disadvantages – all ages to 30, all partnership contexts– IRRs 1.11 to 1.19

• Six ‘age-threshold’ childhood disadvantages– IRRs 1.13 to 2.47

• Eight ‘less favourable’ partnership context terms– IRRs 1.17 to 1.81, (4>=1.40)

• Parsimony – 20 child antecedent & 12 T&C terms (of 617 dummies!)

‘Pervasive’ childhood disadvantages – all ages to 30, all

contexts, both genders and cohorts

– Social housing (1.16), – Any young parent (1.18), – low parental interest (1.29), – fairly ASB (1.11), – <2 high Q tests (1.15)– 2/3 Low Q tests (1.11)

‘Age-threshold’ childhood disadvantages

• Mainly ‘long reach’ to age 25; half are gendered– 16-24:

• clear SES (1.13), • strong SES (1.20)

– 16-24 Females: • any ASB (1.13), • <2 High quartile tests (1.33)

– 16-22: • any Low Q test (1.17)

– 16-19 Females• Very ASB (2.47)

‘Less favourable’ partnership context terms

• Main contrasts by married vs not married– Not married:

• any SES (1.21), • no very interested (1.19), • any ASB (1.17), • <2 High Q test (1.51)• any Low Q test (1.36)

– Not in partnership: • any family disruption (1.49),

– Not in partnership Females: • Strong SES (1.43)

– Never partnered 1970 Cohort: • Social housing (1.81)

Gender & Cohort terms

• GENDER – all excess female legacies of disadvantage; predominantly timing– 16-24, Females:

• any ASB (1.13), • <2 High quartile tests (1.33)

– 16-19, Females:• Very ASB (2.47)

– Not in partnership, Females: • Strong SES (1.43)

• COHORT– Never partnered 1970 Cohort:

• Social housing (1.81)

IRRs for CM Parent’s ages

Father >25 &

Mother >23

Either or both parents ‘young’

1.00 1.18

Net IRRs - Tenure

NCDS BCS70

All Ever Partnered

Never Partnered

No Local Authority

1.00 1.00 1.00

Any Local Authority

1.16 1.16 2.10

Net IRRs – Family Disruption

Not in partnership

In partnership

No Disruption 1.00 1.00

Any Disruption 1.49 1.00

Net IRRs – Parental Interest

Not Married Married

Any very 1.00 1.00

No very, but no little or no

1.19 1.00

Any little or no 1.53 1.29

Net IRRs for SESSES Deprivation

16-24 Not Marr

25-29 Not Marr

16-24 Marr

25-29 Marr

No evidence 1.00 1.00 1.00 1.00

Slight/ Some 1.21 1.21 1.00 1.00

Clear 1.37 1.21 1.13 1.00

Strong 1.64 1.21 1.35 1.00

16-24 Not Part

25-29 Not Part

Strong Female

2.35 1.74

Net IRRs for ASB

ASB Not Married Married

16-19 Fem

20-24 Fem

Men 25+ F

16-19 Fem

20-24 Fem

Men 25+ F

None 1.00 1.00 1.00 1.00 1.00 1.00

Slight 1.33 1.33 1.17 1.13 1.13 1.00

Fairly 1.47 1.47 1.30 1.25 1.25 1.11

Very 3.63 1.47 1.30 3.10 1.25 1.11

Net IRRs - Test Scores

Not Married Married

16-22 23-24 25-29 16-22 23-24 25-29

2/3 Hi 1.00 1.00 1.00 1.00

No Lo 1.74 1.74 1.15 1.15

Any L 2.77 2.37 1.35 1.15

2/3 Lo 3.07 2.63 1.49 1.28

No Lo 2.32 1.74 1.53 1.15

Any L 3.69 3.16 1.79 1.15

2/3 Lo 4.08 3.50 1.98 1.28

Summary 1

• CONSISTENCY of ‘legacies’ of childhood disadvantage

• Childhood antecedents ‘play’ through partnership context slightly more often and more strongly than through age

• Childhood antecedents generally have ‘long age reach’ covering 16-24 or 16-29

• Childhood disadvantage especially associated with excess first birth rates outside of marriage and only slightly more out of co-residential partnership

Gender

• Most child disadvantages apply equally to men & women’s first birth rates (16)

• Excess legacies of disadvantage only for women and link more to timing (3) than context (1)

• Four (of 12) ‘structural’ parameters capture remaining gender differences

Cohort

• Remarkably only one interplay of childhood disadvantage with cohort identified – social housing

• There is also only one ‘structural’ element for cohort, with an IRR of 1.31 for the ‘not married’ group in the 1970 cohort, reflecting the rise in first births outside marriage (both cohabiting and not in partnership)

Limitations

• Further paper needed to explore links of childhood disadvantages, gender and cohort to partnership contexts

• Also correlated errors and multistate multiprocess models?

Conclusion

• Childhood disadvantages are ‘drivers’ of ‘risky’ demographic behaviours: – out-of-wedlock (or out-of-partnership)– Youthful parenthood?

• Important consequences (not documented here):– Partnership instability– Less father involvement– Socioeconomic disadvantage– Poorer mental health– Gender inequities – penalties of risky parenthood

greater for mothers than fathers