Table of Contents | Previous | Next |
C.3 CONSTRUCTION OF TIMELINES
The employment- and education-related outcome variables were constructed from weekly timelines signifying whether the primary caregiver was employed or in a school or training program in each week during the 15 months after random assignment. Similarly, the welfare-related outcome variables were constructed using monthly timelines signifying whether the family was receiving various forms of public assistance benefits in each month. These timelines were constructed using data from the 6- and 15-month Parent Service Interviews.
Timelines were constructed using start and end dates of spells. Positive integers were used to signify that the caregiver was in a spell in a week (month) after random assignment. If the reported day that a spell started or ended was missing, we set the day to "15." However, if the month or year was missing, the relevant timeline entries were set to "missing" using alphabetic codes. A timeline entry could have multiple codes pertaining to overlapping spells. For example, a code of '1B' signified that the caregiver was working on the first job reported in the survey, but also that we were unsure whether she was working on job 2.
The variables pertaining to weeks spent employed, in school or training, or on welfare during the 15 months after random assignment were constructed by summing the number of weeks (months) that the relevant timelines had positive codes. The variables were set to zero if the family had no spells, and they were set to "missing" if any timeline entry had a missing code but no positive code. Similarly, variables pertaining to hours spent in employment and education activities were constructed using the timelines and survey information on the number of hours per week the caregiver usually spent in each activity. Finally, we constructed variables pertaining to the amount of public assistance benefits that were received using the welfare timelines and information on the monthly amount of benefits received for each spell of receipt.
| Table of Contents | Previous | Next |