California Health Interview Survey

 

 Join our Newsletter

 

Join Our Newsletter

A monthly e-mail of breaking news, data, and publications from the Center.

Join
print share

CHIS Data Quality & the Survey Environment

CHIS provides high quality data that accurately represent the California household population. Since CHIS 2001, the survey has been guided by the data quality objectives of the Total Survey Error perspective (Groves et al., Survey Methodology), an approach that collectively addresses multiple threats to survey quality. 

Like all surveys, CHIS is challenged by a rapidly changing survey environment in which people are harder to reach and can be selective of about how and when they are reached.  Response rates in nearly all population-based surveys have declined over the past few decades,  and the recent growth of cellular telephones introduces potential coverage problems for traditional random digit dial (RDD) sampling methods. Declining response rates may increase the potential for nonresponse bias in CHIS estimates, and the growth of cellular telephones may increase the potential for noncoverage bias in CHIS estimates. View our short bibliography of research on these data quality issues. 

See some of things that CHIS does to produce high-quality data each cycle​

CHIS is at the forefront of efforts to scientifically understand and address potential nonresponse and noncoverage bias. Below are studies we have done to assess methodological issues and guide our future efforts to maintain the highest standards for data quality. While no single study can definitively demonstrate data quality, the multiple studies taken as a whole consistently point in the same direction: CHIS data are high quality and accurately represent California's household population.​ 

See some of our data quality assessments below. 

Assessing and Addressing Potential Noncoverage Bias​

CHIS 2003 Comparison of CHIS Medicaid Estimates with Other Surveys

CHIS 2003 Health Survey Benchmarking Study

​CHIS 2007 Nonresponse Analysis​

Using Neighborhood Characteristics to Assess Nonresponse Bias​​