Raising the Bar: Impacts and Implementation of the New Heights Program for Expectant and Parenting Teens in Washington, dc

Table B.1. Data sources and key data elements for file linkage and analysis 
Data source 
Purpose
Data elements for file 
linkage 
Data elements for analysis 
DCPS 
Student records 
Identify sample; 
obtain demographic 
records; obtain 
outcome information 
DCPS student identifier 
Name 
Birth date 
Sex 
Home address 
Covariates 
•
Age 
•
Race/ethnicity 
Outcomes 
•
High school days attended
•
High school excused 
absences 
•
High school unexcused 
absences 
•
High school credits earned 
•
High school graduation 
School admissions 
and withdrawal 
records 
Identify enrollment in 
a study school 
DCPS student identifier 
Enrollment in a study school 
DC DOH 
Vital records 
Identify teens who 
gave birth in D.C.
Name 
Maiden name 
Mother’s birthdate 
Home address 
Covariates 
•
Measure of having given birth 
•
Child’s age 
DC DHS 
New Heights 
participation records 
Identify treatment 
group of New Heights 
participants 
DCPS student identifier 
Participation in New Heights 
program 
Accounting for changes in record keeping and measurement 
This evaluation examines the impact of New Heights on attendance, credits earned, and high 
school graduation of parenting females. Each of these outcome variables was created from DCPS 
data (Table B.1). Through examining the data and discussing them with DCPS staff, we 
identified some changes in the ways DCPS recorded and measured these outcomes over the eight 
years the records span.
DCPS changed the cutoff used in defining whether a student attended a day of school in the 
2013–2014 school year. Previously, students who attended at least 60 percent of the day were 
counted as attending a full day. The new cutoff states that students must attend 80 percent of the 
day to be counted as attending a full day.
If this change affects parenting and nonparenting youth similarly, then our analysis will help 
adjust for the changes. However, this definitional change does have the potential to affect 
parenting youth differently from nonparenting youth, for example, if parenting youth were more 
likely to attend school for 60 to 80 percent of a day. Such a circumstance could lead to a negative 
bias in the estimate of New Heights’ impact on attendance. However, we have concluded that 
this change does not significantly affect the credibility of our main findings because (1) it is a 
relatively nuanced definitional change, (2) our difference-in-difference estimation approach 
partially adjusts for this, and (3) a negative bias enhances the credibility of our positive impact 
finding. 
B.5 

---

APPENDIX B: ADMINISTRATIVE DATA PREPARATION
 
Another change included the number of credits students could earn in a semester, which 
increased after the expansion of New Heights. DCPS staff speculate that reasons for this include 
better scheduling and higher graduation requirements. If the definition change affects parenting 
youth differently from nonparenting youth, then it could bias the estimate of New Heights’ 
impact on number of credits earned. It is less clear whether this would be a positive or negative 
bias. However, our difference-in-difference estimation approach does partially adjust for this.