Tracking children in the largest school district in the United States: A comparison of administrative data and field tracking from preschool to kindergarten
Jul 26, 09:00
Maintaining contact with longitudinal study participants is essential to the success of such studies. To reduce nonresponse and associated loss of data quality, studies often include a variety of tracking activities in their data collection plans to keep track of respondents over time. This paper will examine the results and effectiveness of an intensive tracking plan employed on Making Pre-K Count (MPC). MPC used both field tracking methods, as well as school administrative data to track 2,700 New York City children from preschool into kindergarten (Spring 2015 into Fall 2015).
MPC was a random assignment evaluation of a preschool mathematics curriculum, with treatment and control assignment conducted at the school level. Specifically, the study rigorously evaluated the effect of a mathematics-focused curriculum called Building Blocks, in 69 preschools and over 170 classrooms in New York City (NYC). Thirty-five preschools were assigned to receive the math curriculum over two years, while the comparison group proceeded as usual. Outcomes for children were assessed during both their preschool and kindergarten years.
MPC employed a variety of methods to track children from preschool to kindergarten: (1) collection of enrollment plans via preschool rosters and parents; (2) in the field by a team of data collectors, and (3) analysis of Fall 2015 kindergarten enrollment administrative data maintained by the school district. For a portion of the sample (approx. 1,700 children), direct child assessments were conducted at the end of the kindergarten year (Spring 2016). These assessments allowed definite confirmation of enrollment, versus relying on the accuracy of administrative data.
Analysis of administrative data may provide a less costly approach to tracking than does the approach of hiring, training, and supervising data collectors. We will examine the degree to which public school administrative data can be used to effectively track a sample of low-income, urban children from preschool to kindergarten in the United States’ largest school district. This will be achieved through comparisons of the enrollment data collected via each of the methods noted above. We will also investigate how other traditional tracking methods, such as the collection of enrollment plans directly from parents, compare with administrative enrollment data. For those children not successfully tracked via administrative data, we will examine key demographic characteristics, such as primary household language, to discern if there is an identifiable pattern of missingness using the administrative data approach.