ID Variables

Two systems of ID variables have been employed in CHNS data files: the original system used in surveys before 2004, and a revised system used thereafter. All old IDs have been changed to new IDs in all files.

Household ID (HHID)

HHID is a nine-digit numeric variable that uniquely identified each household that had been seven digits in the old cross-sectional files. The variables T1 through T5 (documented in the questionnaire) were concatenated to form the HHID. Each HHID value represented one household. When the unit of analysis for a file/table was household, 1 HHID = 1 row = 1 household = 1 observation. Observations in these files were sorted by HHID and survey year (Wave). Thus the key sort variables for these files were HHID and Wave.

Individual ID (IDind)

The IDind is a twelve-digit numeric variable that uniquely identified for all participants. Each participant will have the same ID in all datasets and in all survey years. The unique ID will not change over time and will facilitate data merges across datasets and survey years. Each IDind value represented one individual. When the unit of analysis for a file/table was individual, 1 IDind = 1 row = 1 individual = 1 observation. Observations in these files are sorted by IDind, and Wave, i.e., the key sort variables were IDind, and Wave.

Community ID (COMMID)

A third ID variable, COMMID, is a six-digit numeric variable that uniquely identified each community. The variables T1 through T4 were concatenated to create COMMID. Each COMMID value represented one community. When the unit of analysis for a file was community, 1 COMMID = 1 community = 1 observation. Observations in these files were sorted by COMMID and Wave, i.e., the key sort variables were COMMID and Wave. Although COMMID was not required for most file merges, this variable was included on all data sets to facilitate merges with community-level files.

When the unit of analysis was something other than individual, household or community (e.g., job, livestock type, food item, health facility), a variable that identified this unit was included on the file (e.g., JOB, F11, FOODCODE, Q1). For the files/tables where job was the unit of analysis, for example, each value of the variable JOB represented one occupation. That is, 1 JOB = 1 row = 1 occupation = 1 observation. Observations in these files were sorted by HHID, LINE, and JOB, i.e., the key sort variables were HHID, LINE, and JOB.

Survey Year (WAVE)

The WAVE is a four-digit numeric variable that identified survey year (i.e, 1989, 1991, 1993, 1997, 2000, 2004, 2006, 2009, 2011, 2015). This variable was used in master longitudinal files only.