12.07.2015 Views

Fourth National Incidence Study of Child Abuse and Neglect (NIS–4)

Fourth National Incidence Study of Child Abuse and Neglect (NIS–4)

Fourth National Incidence Study of Child Abuse and Neglect (NIS–4)

SHOW MORE
SHOW LESS
  • No tags were found...

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

UNDUPLICATION NIS provides estimates <strong>of</strong> the numbers <strong>of</strong> maltreated children, so it is critical for thestudy to avoid counting the same child more than once. The purpose <strong>of</strong> unduplication is toidentify children who enter the study data on multiple data forms <strong>and</strong> reduce their information toa single record for analysis. Unduplicating NIS–4 data consisted <strong>of</strong> three main steps:• Identifying child-level records that may be duplicates (c<strong>and</strong>idate pairs)• Deciding whether the c<strong>and</strong>idate pair records were true duplicates• Unifying duplicate recordsC<strong>and</strong>idate pairs. Matches on subsets <strong>of</strong> 8 key data items helped to identifyc<strong>and</strong>idate pairs:First name AgeLast name initial Ethnicity/raceGenderCity <strong>of</strong> residenceDate <strong>of</strong> birth Number <strong>of</strong> children in householdC<strong>and</strong>idate pairs for child-level records from CPS Maltreatment <strong>and</strong> Sentinel dataforms were identified 3 ways:• Manually using a computerized sorting system• Using the NIS–3 rule-based algorithm that identified 2 <strong>of</strong> 3 matching patterns• Using a probability-based matching s<strong>of</strong>tware designed to identify matchesDuring the first stage, all data forms in 9 small counties were examined manually.This entailed unduplication staff sorting all the child-level records in each county by various keydata items <strong>and</strong> flagging pairs <strong>of</strong> records that appeared to be potential duplicates. Statisticiansused the data in these counties <strong>and</strong> these initial c<strong>and</strong>idate-duplicate decisions to guide thesettings <strong>of</strong> parameters on the probability-based matching s<strong>of</strong>tware so that it would, as closely aspossible, identify the same c<strong>and</strong>idate pairs. In addition, the unduplication task leader adjustedthe NIS–3 rule-based algorithm so that it would not generate numerous false-positive c<strong>and</strong>idates.Once these preparations were completed, the adjusted NIS–3 rule-based algorithm <strong>and</strong> theA-23

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!