Table 3 Random effects meta-analysis of regression of daily cigarette consumption (binary) on childhood predictors: Total sample (N = 202,898)

From: A cross-national analysis of childhood predictors of daily smoking in adulthood

 

Estimated proportion of effects by threshold

 

Global p-value

Variable

Category

RR

95% CI

<0.90

>1.10

I2

Relationship with mother

(Ref: very bad/somewhat bad)

     

0.045*

Very good/somewhat good

0.94

(0.90,0.99)

0.00

0.00

<0.1a

 

Relationship with father

(Ref: very bad/somewhat bad)

     

0.165

Very good/somewhat good

0.94

(0.89,0.99)

0.27

0.00

31.0

 

Parental marital status

(Ref: parents married)

     

6.99e-06**

Divorced

1.26

(1.20,1.33)

0.00

1.00

32.7

 

Single, never married

1.16

(1.06,1.27)

0.00

0.64

62.1

 

One or both parents had died

1.10

(1.03,1.17)

0.00

0.50

8.4

 

Subjective financial status of family growing up

(Ref: got by)

     

0.002**

Lived comfortably

1.00

(0.97,1.04)

0.00

0.00

29.4

 

Found it difficult

0.99

(0.94,1.03)

0.05

0.05

29.6

 

Found it very difficult

1.06

(0.99,1.13)

0.05

0.27

16.5

 

Abuse

(Ref: no)

     

2.35e-05**

Yes

1.26

(1.19,1.34)

0.00

0.95

64.1

 

Outsider growing up

(Ref: no)

     

1.54e-05**

Yes

1.20

(1.14,1.28)

0.00

0.82

59.8

 

Self-rated health growing up

(Ref: good)

     

1.10e-05**

Excellent

1.03

(0.98,1.10)

0.09

0.27

60.8

 

Very good

0.99

(0.95,1.04)

0.09

0.05

42.2

 

Fair

0.97

(0.92,1.02)

0.05

0.00

12.5

 

Poor

0.52

(0.17,1.55)

0.45

0.23

99.4

 

Immigration status

(Ref: born in this country)

     

1.77e-05**

Born in another country

0.90

(0.75,1.10)

0.50

0.23

85.7

 

Age 12 religious service attendance

(Ref: never)

     

2.19e-05**

At least 1/week

0.96

(0.90,1.03)

0.32

0.14

56.3

 

1–3/month

1.05

(0.95,1.16)

0.14

0.36

75.7

 

<1/month

1.00

(0.95,1.05)

0.05

0.05

30.5

 

Age; year of birth

(Ref: age 18–24; 1998–2005)

     

1.44e-06**

age 25–34; 1988–1998

1.39

(1.27,1.53)

0.00

0.95

73.6

 

age 35–44; 1978–1988

1.54

(1.36,1.75)

0.00

0.95

85.8

 

age 45–54; 1968–1978

1.51

(1.31,1.74)

0.00

0.95

87.3

 

age 55–64; 1958–1968

1.53

(1.30,1.80)

0.00

0.86

87.8

 

age 65–74; 1948–1958

1.14

(0.92,1.42)

0.32

0.45

87.8

 

age 75–84; 1938–1948

0.22

(0.04,1.29)

0.64

0.18

99.5

 

85 or older; 1938 or earlierb

0.04

(0.00,0.37)

0.68

0.27

99.5

 

Gender

(Ref: male)

     

1.53e-06**

Female

0.32

(0.19,0.52)

0.86

0.00

99.7

 

Otherb

0.11

(0.01,1.05)

0.61

0.28

99.4

 
  1. * p < 0.05; ** p < 0.004 (Bonferroni corrected threshold).
  2. aEstimate of heterogeneity is likely unstable. See our online Supplementary Figs. for more detail on the heterogeneity of effects.
  3. bGroup is very small (<0.1% of the observed sample) within several countries leading large uncertainty in this estimate or even complete separation—be cautious about interpreting this estimate; CI = confidence interval; the Estimated proportion of effects is the estimated proportion of effects above (or below) a threshold based on the calibrated effect sizes41; I2 is an estimate of the variability in means due to heterogeneity across countries vs. sampling variability; the Global p-value corresponds to the joint test of the null hypothesis that the country-specific joint parameter Wald tests (all parameters within variable groups are zero) are all null in all 22 countries; and additional details of heterogeneity of effects are available in our online Supplementary Figs.
Back to article page