Title: The Effect of Older Siblings on Adolescent Behavior: Is It Siblings or is it Parents
1The Effect of Older Siblings on Adolescent
Behavior Is It Siblings or is it Parents?
Laura Argys Department of Economics University of
Colorado at Denver Susan Averett Department of
Economics and Business Lafayette College Daniel
Rees Department of Economics University of
Colorado at Denver
2Birth Order and Parental Supervision
Price (2005) Only study to examine parental time
and birth order Uses data from American Time Use
Survey Finds that parents spend 20-30 minutes
less per day with later-born children.
3Parental Supervision and Child Behaviors
Parental (Maternal) work and cognitive outcomes
in young children Leibowitz (1977), Blau and
Grossberg (1992), Brooks-Gunn, Han and Waldfogel
(2002) Parental supervision and risky
behaviors substance use Richardson et al.
(1993) behavior problems Vandall and Ramanan
(1991) substance use, theft, truancy Aizer
(2004)
4Research Objectives
- To confirm the existence of a birth-order effect
on risky behaviors. - To disentangle the effect of
- Parents
- differential supervision of children by
birth-order - Siblings
- initiation
- modeling
5Previous Empirical Findings on Economic Outcomes
Birth order and Education Olneck and Bill,
1979 No Blake, 1981 No Hauser and
Sewell, 1983 No Behrman and Taubman,
1986 Yes Birth order and Earnings Olneck and
Bill, 1979 No Behrman and Taubman,
1986 No Kessler, 1991 No Sex
Composition Butcher and Case,
1994 Yes Kaestner, 1997 No Hauser and
Kuo, 1998 No
6Previous Empirical Findings on Other Outcomes
Birth order on age at first intercourse
Rodgers st al., 1992 Older sibling graduation
on younger siblings graduation Oettinger,
2000 Older brother characteristics on drug
use Brook et al., 1991 Older sibling
sexual behavior on sexual initiation Widmer,
1997 Haurin and Mott, 1990 Older sibling
behavior on substance use Ouyang, 2004
7Model
An individuals propensity to engage in a
particular risky activity
A dichotomous variable indicating an older
sibling
A vector of controls for family size
Other individual and family characteristics
8Figure 1 Older siblings and substance
use and sexual behavior
9Figure 2 Older siblings and criminal
and delinquent activities
10Data
National Longitudinal Survey of Adolescent Health
Over 20,000 adolescents In grades 7-12 in
1995 Household roster and family
structure Risk-taking activities Family
background Parental Supervision
11Conclusions
The existence of an older sibling Increases the
probability of cigarette, alcohol and marijuana
use and sexual activity Has no consistent effect
on younger siblings delinquent activities Parenta
l Supervision does not explain the older sibling
effect Parental supervision reduces risky
behaviors . But the effect of older siblings
remains.
12Table 1. Sample Means of Child Outcomes, by
Gender and the Presence of an Older Sibling
13Table 2. Sample Means, by Gender and the
Presence of an Older Sibling
Standard deviations are shown in parentheses.
14Table 2. Sample Means, by Gender and the
Presence of an Older Sibling (Continued)
Standard deviations are shown in parentheses.
15Table 3a. Determinants of Substance Use and
Sexual Activity Females
16Table 3a. Determinants of Substance Use and
Sexual Activity Females (Continued)
Standard errors are shown in parentheses.
Statistically significant at the 99, 90,
and 90 confidence level. Controls also include
dichotomous variables identifying missing values
for categorical variables.
17Table 3b. Determinants of Substance Use and
Sexual Activity Males
18Table 3b. Determinants of Substance Use and
Sexual Activity Males (Continued)
Standard errors are shown in parentheses.
Statistically significant at the 99, 90,
and 90 confidence level. Controls also include
dichotomous variables identifying missing values
for categorical variables.
19Table 4a. Determinants of Delinquent Behaviors
Females
20Table 4a. Determinants of Delinquent Behaviors
Females (Continued)
Standard errors are shown in parentheses.
Statistically significant at the 99, 90,
and 90 confidence level. Controls also include
dichotomous variables identifying missing values
for categorical variables.
21Table 4b. Determinants of Delinquent Behaviors
Males
22Table 4b. Determinants of Delinquent Behaviors
Males (Continued)
Standard errors are shown in parentheses.
Statistically significant at the 99, 90,
and 90 confidence level. Controls also include
dichotomous variables identifying missing values
for categorical variables.
23Table 5a. Proportion Reporting Parental
Supervision, by Gender, Family Size and the
Presence of an Older Sibling
24Table 5a. Proportion Reporting Parental
Supervision, by Gender, Family Size and the
Presence of an Older Sibling Two
Child Family
25Table 5b. Regressions of Parental Supervision
on Family Structure
Standard errors are shown in parentheses.
Statistically significant at the 99, 90,
and 90 confidence level.
26Table 6. Determinants of Substance Use and Sexual
Activity, Selected Marginal Probabilities
Standard errors are shown in parentheses.
Statistically significant at the 99, 90,
and 90 confidence level. Tables control for the
same variables that appear in tables 3a. and 3b.
Controls also include dichotomous variables
identifying missing values for categorical
variables.
27Table 7. Determinants of Delinquent Behaviors,
Selected Marginal Probabilities
Standard errors are shown in parentheses.
Statistically significant at the 99, 90,
and 90 confidence level. Tables control for the
same variables that appear in tables 4a. and 4b.
Controls also include dichotomous variables
identifying missing values for categorical
variables.
28Table 8. Regressions of Parental Supervision on
Sibling Gender and Family Structure
29Table 9a. Effects of Siblings on Substance Use
and Sexual Activity, Selected Marginal-
Probabilities Females
Standard errors are shown in parentheses.
Statistically significant at the 99, 90,
and 90 confidence level. Tables control for the
same variables that appear in tables 3a. and 3b.
Controls also include dichotomous variables
identifying missing values for categorical
variables.
30Table 9b. Effects of Siblings on Substance Use
and Sexual Activity, Selected Marginal-
Probabilities Males
Standard errors are shown in parentheses.
Statistically significant at the 99, 90,
and 90 confidence level. Tables control for the
same variables that appear in tables 3a. and 3b.
Controls also include dichotomous variables
identifying missing values for categorical
variables.