Estimated Effects of Teen Birth on Various Outcomes for Alternative and Placebo Treatment Groups
| Placebo: Pre-Scores for Teen Aunts/Uncles (t =-2) (IND+FAM) (1) | Placebo: Nonchildbearing Families (IND+FAM) (2) | Placebo: Nonchildbearing Families (IND+FAM+NBHD) (3) | Alternative: Older Siblings as Treated (IND+FAM) (4) | |
|---|---|---|---|---|
| Test scores | 0.321 | 0.199 | −0.614 | |
| (1.111) | (1.085) | (1.191) | ||
| N | 1,009 | 1,140 | 990 | |
| Repeats grade | 0.704 | −5.216+ | −5.119+ | |
| (0.954) | (2.917) | (3.044) | ||
| N | 1,078 | 1,696 | 1,392 | |
| Drops out | −0.496 | −0.130 | ||
| (2.351) | (2.292) | |||
| N | 1696 | 1392 | ||
| Juvenile justice | −0.106 | −1.470 | −3.474+ | |
| (0.108) | (1.628) | (1.801) | ||
| N | 1,078 | 1,696 | 1,392 | |
| Ever attends any college | 1.857 | 0.980 | −25.530+ | |
| (3.839) | (4.121) | (14.174) | ||
| N | 1,174 | 941 | 80 | |
| Obtains any degree or certificate | 1.209 | −0.674 | 17.431 | |
| (3.261) | (3.446) | (10.750) | ||
| N | 1,174 | 941 | 80 | |
| Obtains a four-year degree | −0.240 | 0.500 | 5.731 | |
| (2.719) | (2.889) | (11.346) | ||
| N | 1,174 | 941 | 80 |
Notes: Robust standard errors clustered by family ID. All analyses based on the trajectory matches with other controls and same neighborhood requirement. Column 1 conducts the analysis for the older siblings of teen mothers. These teen aunts/uncles are over 18 years old. I use an IND+FAM match because the IND+FAM+NBHD model has very low N. I use the six, seven, and eight years before the birth to estimate the trajectory, as these are the years with at least 50 percent of the test score data for the siblings who are 18 or older at the time of the birth. Outcomes are postsecondary, though it is possible that the siblings may have started or completed college by the time of birth. The outcome is Column 2 examines the outcomes for the IND+FAM control group two years before the birth. Columns 3 and 4 conduct the IND+FAM and IND+FAM+NBHD analysis using 200 randomly selected families without teen childbearing in their family who had been used in the IND+FAM analysis as controls. These former controls are treated as a placebo group and matched to a new set of matches, and the analysis displays the comparison between these placebos and their matched controls.