Longitudinal Data Structure - Economics Department - University of Dallas

Longitudinal Data Structure - Economics Department - University of Dallas

CARE_CCS

Data collection is opportunistic. A new record is generated each time a household visits Crossroads Community Services to receive food. Multiple records for each household can be linked to create a longitudinal/panel data structure. On May 31st each year, an annual set of data files is archived in the CCL Database. All data is collected based on responses given by the household member who comes to pick up food. We refer to this household member as the "household reporter". Each downloaded set of data contains 5 distinct files which are linkable at the household level:

  • Contact File--this file includes socio-demographic and employment information about the household reporter and household-level information. There is one observation per household and this information is updated every 6-12 months. When possible, information contained in this file is validated using external documents.
  • Household File--this file contains socio-demographic and employment information about every household member, not including the household reporter. There is one observation per household and this information is updated every 6-12 months. To the extent possible, information contained in this file is validated using external documents (i.e. school records for child age, etc.).
  • Longitudinal File--this file contains responses to many self-report survey questions and some measured outcomes. The longitudinal file includes all constructs measured more frequently than every 6-12 months. This file contains the full set of responses to each construct, thus there is more than one observation per household. Many constructs are measured at every visit (maximum frequency is "monthly"), and others are collected on a rotating basis. Figure 2 below describes the number of records per household in the Longitudinal File.
  • Food File--this file details the food selections of each household at each visit as well as the foods that were available for selection. This data is collected at every visit (maximum frequency is "monthly").
  • Geoidentifiers--this file contains the residential address of households and is highly confidential. Access to this file is restricted unless researchers can prove the ability to maintain confidentiality and obtain approval from their Institutional Review Board. There is one observation per household, and this information is updated every 6-12 months. Addresses are validated using rental agreements or utility bills.

Figure 2. Number of Visits per Household (May 2013-April 2017)

ccld_visits

The CCL Database includes data about low-income households in Dallas County who receive food from Crossroads Community Services Walk-in Pantry or one of many Community Distribution Partners (CDP). CDPs consist of a variety of organizations that obtain food packages from Crossroads to re-distribute to households in the communities they serve. Prior to December 2016, data collection at CDPs was done on paper and only included those constructs collected annually. Beginning in 2016, select CDPs began collecting the full set of constructs using a tablet-based ClientCare data collection tool.

Table 2. Implementation of Data Collection at Community Distribution Partner Sites

CDP

Figure 3. Frequency of Households Represented in the CCL Database by Zip Code (April 2013-May 2017)

 ccld_map