Online Resource 2
Investing in Children: Changes in Parental Spending on Children, 1972–2007
Sabino Kornrich and Frank Furstenberg
Effects of Children’s Gender Composition on Spending in Multivariate Models
The following regressions show the effects of the gender composition of children on spending on each of the three categories of spending. We present these results to illustrate that significant differences in means,based on ttests, are robust to the inclusion of controls for household characteristics thatinfluence spending. We examine spending across three types of gender composition: households with only girls, those with mixed-gender children, and households with only boys. We expect that households with mixed-gender children will display levels of spending between households with only boys and those with only girls.We show results separately by specific goods to match our analyses using ttests. Each table shows regression results and levels of significance for one of the time periods we investigate. Thus, significance tests reflect the rejection of the null hypothesis that a coefficient is zero within each time period.
These results provide confirmation of significant differences in spending by child gender: there are significant effects of being in a household with only girls, compared with only boys, in many cases. Also, households with children of both sexes spend amounts that are intermediate between households with only girls and households with only boys, supporting the notion that the gender of children is closely tied to parental investment in the form of spending.
Specific results show that in 1972–1973, parents of only girls spent significantly less on education than parents of only boys. Those with both boys and girls also spent less on education than households with only boys, but the coefficient is not as large as that for households with only girls, showing that gender also has effects in this case even without the extreme comparison of only girls and only boys. In the same year, households with girls and households with children of both sexes were more likely to spend on a variety of goods,such asclothes, toys, and games. Thus, in the early 1970s, parents invested substantially more in boys’ education, but they invested more in the use of goods for girls.
The link between children’s gender composition and spending changed substantially over the subsequenttime periods. In both the 1980s and 1990s, there wereno significant differences in spending. However, by the last data point in the late 2000s, as in ttests, households with only female children present spent significantly more on education than other households. These results thus show a substantial shift in the importance of the gender of children over time. As girls have become more likely to pursue higher education and to pursue long careers in the labor market, parents are investing more, even net of a variety of differences that could exist between households with only male children, only female children, and children of both sexes.
Table S1: Regression results, 1972-3 data, separately by good, showing effect of child gender on spending.
Variable / Education / Child Care / GoodsIntercept / -2053.7 / *** / 55.4 / 267.4 / *
Age of youngest child / 29.7 / ** / -33.1 / *** / 70.7 / ***
Age of youngest child squared / 1.2 / ** / -0.3 / * / -4.6 / ***
Proportion of earnings from wife / 47.1 / 353.2 / *** / 7.9
Wife works part time / -95.1 / * / 82.2 / *** / 6.2
Wife works full time / -49.4 / 174.2 / *** / 16.9
Parental education (no high school degree is reference category)
High school graduate / 68.6 / 41.5 / ** / 15.5
Some college / 141.9 / ** / 71.5 / *** / 94.5 / ***
College degree / 581.2 / *** / 97.5 / *** / 127.7 / ***
Family structure (two-parent family is reference category)
Single mother / 96.7 / 335.1 / *** / 89.6 / ***
Single father / -59.0 / 279.8 / *** / 19.0
Other families / -299.3 / 209.8 / *** / -9.6
Gender of children (boys only is reference category)
Only girls / -228.5 / *** / -3.8 / 103.4 / ***
Mixed gender / -124.5 / * / 2.7 / 64.2 / **
Number of children (four or more is reference category)
One / 18.5 / 510.3 / *** / 487.5 / ***
Two / 63.1 / 228.0 / *** / 206.4 / ***
Three / 35.4 / 107.3 / *** / 86.8 / ***
Household income in 1000s of dollars / 23.0 / *** / 1.4 / + / .7
Earnings decile (top decile is reference category)
1 / 1561.2 / *** / -180.6 / + / -365.0 / **
2 / 1325.1 / *** / -66.8 / -315.0 / **
3 / 1152.2 / *** / -61.2 / -251.2 / **
4 / 1016.6 / *** / -91.1 / -209.1 / *
5 / 885.0 / *** / -46.4 / -220.3 / **
6 / 620.1 / ** / -20.9 / -159.8 / *
7 / 589.8 / ** / -30.3 / -122.3 / +
8 / 386.0 / * / -14.7 / -75.0
9 / 305.4 / * / 22.9 / -42.8
Notes: +:p<.1, *: p<.05, **:p<.01, ***:p<.001
Table S2: Regression results, 1983-4 data, separately by good, showing effect of child gender on spending.
Variable / Education / Child Care / GoodsIntercept / -2731.9 / *** / 269.1 / + / 416.8 / ***
Age of youngest child / 78.2 / *** / -83.4 / *** / 18.7 / ***
Age of youngest child squared / -.6 / 1.4 / *** / -2.4 / ***
Proportion of earnings from wife / -26.6 / 354.1 / *** / 49.5 / +
Wife works part time / 144.9 / * / 181.2 / *** / -2.6
Wife works full time / 24.7 / 490.4 / *** / -11.3
Parental education (no high school degree is reference category)
High school graduate / 228.2 / *** / 2.3 / 79.1 / ***
Some college / 484.4 / *** / 83.3 / * / 108.0 / ***
College degree / 1216.2 / *** / 186.7 / *** / 214.1 / ***
Family structure (two-parent family is reference category)
Single mother / -244.1 / ** / 226.8 / *** / -36.1
Single father / -294.5 / 485.8 / *** / 21.1
Other families / -349.7 / *** / 137.4 / ** / -79.4 / **
Gender of children (boys only is reference category)
Only girls / 95.5 / -26.1 / 14.7
Mixed gender / 8.1 / -5.2 / 9.7
Number of children (four or more is reference category)
One / 76.8 / 643.0 / *** / 586.2 / ***
Two / -65.0 / 333.6 / *** / 277.4 / ***
Three / -45.4 / 140.5 / ** / 114.0 / **
Household income in 1000s of dollars / 22.5 / *** / 1.2 / 1.9 / **
Earnings decile (top decile is reference category)
1 / 2459.2 / *** / -284.2 / * / -322.9 / ***
2 / 2042.2 / *** / -243.8 / + / -340.7 / ***
3 / 1701.8 / *** / -241.1 / + / -309.7 / ***
4 / 1523.4 / *** / -180.9 / -283.6 / ***
5 / 1621.6 / *** / -157.5 / -233.5 / **
6 / 1021.8 / *** / -114.0 / -186.2 / **
7 / 829.0 / *** / -68.4 / -167.1 / **
8 / 596.4 / *** / 89.3 / -90.4
9 / 234.6 / * / 22.2 / -106.9 / *
Notes: +:p<.1, *: p<.05, **:p<.01, ***:p<.001
Table S3: Regression results, 1994-5 data, separately by good, showing effect of child gender on spending.
Variable / Education / Child Care / GoodsIntercept / -2691.9 / *** / 60.1 / 142.7
Age of youngest child / 105.3 / *** / -127.4 / *** / 0.2
Age of youngest child squared / -1.0 / 2.3 / *** / -1.9 / ***
Proportion of earnings from wife / -57.7 / 295.7 / *** / -2.8
Wife works part time / 40.3 / 269.4 / *** / -13.1
Wife works full time / 92.3 / 513.2 / *** / -18.1
Parental education (no high school degree is reference category)
High school graduate / 128.0 / 43.0 / 73.6 / **
Some college / 420.0 / *** / 118.9 / * / 141.0 / ***
College degree / 1113.2 / *** / 235.1 / *** / 168.6 / ***
Family structure (two-parent family is reference category)
Single mother / -210.0 / 351.6 / *** / -15.0
Single father / -340.5 / 452.2 / *** / 32.1
Other families / -434.7 / *** / 238.2 / *** / -79.6 / **
Gender of children (boys only is reference category)
Only girls / 4.2 / 53.1 / 0.6
Mixed gender / 55.1 / 49.4 / 37.4
Number of children (four or more is reference category)
One / 355.6 / * / 919.4 / *** / 707.9 / ***
Two / 27.7 / 508.7 / *** / 333.7 / ***
Three / -107.5 / 232.0 / *** / 165.4 / ***
Household income in 1000s of dollars / 26.4 / *** / 4.9 / *** / 3.9 / ***
Earnings decile (top decile is reference category)
1 / 2050.9 / *** / -212.0 / -25.7
2 / 1741.0 / *** / -210.4 / -5.2
3 / 1541.5 / *** / -171.9 / 6.2
4 / 1253.8 / *** / -118.2 / 39.5
5 / 962.8 / ** / -141.1 / -34.5
6 / 744.7 / * / -74.6 / -1.9
7 / 406.9 / -122.0 / 16.2
8 / 37.4 / -56.7 / 68.8
9 / -17.9 / -81.3 / 69.0
Notes: +:p<.1, *: p<.05, **:p<.01, ***:p<.001
Table S4: Regression results, 2006-7 data, separately by good, showing effect of child gender on spending.
Variable / Education / Child Care / GoodsIntercept / 975.6 / * / 694.2 / *** / 412.4 / ***
Age of youngest child / 113.1 / *** / -177.2 / *** / -21.3 / ***
Age of youngest child squared / -0.5 / 4.1 / *** / -0.7 / ***
Proportion of earnings from wife / -162.9 / 255.5 / *** / -9.0
Wife works part time / 99.1 / 433.6 / *** / 7.2
Wife works full time / -340.3 / * / 631.4 / *** / -75.4 / **
Parental education (no high school degree is reference category)
High school graduate / 201.6 / -11.9 / 39.7 / +
Some college / 481.8 / ** / 120.1 / * / 89.5 / ***
College degree / 1178.5 / *** / 387.3 / *** / 97.4 / ***
Family structure (two-parent family is reference category)
Single mother / -135.5 / 435.3 / *** / 17.6
Single father / -671.3 / * / 549.5 / *** / 58.8
Other families / -659.9 / *** / 274.2 / *** / -23.2
Gender of children (boys only is reference category)
Only girls / 393.9 / *** / 72.6 / 67.3 / ***
Mixed gender / 160.9 / 69.9 / 51.0 / *
Number of children (four or more is reference category)
One / 320.2 / 988.5 / *** / 507.9 / ***
Two / 51.6 / 602.5 / *** / 258.5 / ***
Three / -98.7 / 210.2 / * / 69.5 / *
Household income in 1000s of dollars / 5.3 / *** / 2.9 / *** / 1.3 / ***
Earnings decile (top decile is reference category)
1 / -1702.3 / *** / -679.7 / *** / -227.9 / ***
2 / -1848.4 / *** / -632.2 / *** / -228.2 / ***
3 / -1853.7 / *** / -591.6 / *** / -185.6 / **
4 / -2119.1 / *** / -575.8 / *** / -159.5 / ***
5 / -2098.5 / *** / -643.3 / *** / -178.9 / ***
6 / -1840.7 / *** / -591.1 / *** / -134.2 / **
7 / -2005.6 / *** / -548.3 / *** / -92.7 / +
8 / -1700.4 / *** / -538.9 / *** / -99.5 / *
9 / -1646.3 / *** / -356.6 / ** / -29.2
Notes: +:p<.1, *: p<.05, **:p<.01, ***:p<.001