Survey research methods in evaluation and case-control studies

Area of Science:

• Epidemiology
• Biostatistics
• Survey Methodology

Background:

• Survey research is integral to analytic studies, including intervention evaluation and population-based case-control studies.
• Controlling for confounding is vital in evaluation studies lacking random assignment.
• Population-based controls are increasingly used in case-control studies, but traditional methods face challenges.

Purpose of the Study:

• To provide a comprehensive overview of design and analysis considerations in survey research for analytic studies.
• To discuss methods for confounder control in evaluation studies.
• To explore challenges and alternatives in selecting controls for case-control studies.

Main Methods:

• Overview of design strategies, including matching for confounder control.
• Discussion of analytic techniques such as propensity scoring, relating it to survey weighting.
• Examination of alternative sampling frames for control selection, moving beyond random digit dialling (RDD).

Main Results:

• Propensity scoring offers a robust method for confounder adjustment, similar to survey weighting.
• Random digit dialling (RDD) surveys face declining response rates, necessitating alternative control identification methods.
• Survey sampling methods and variance estimation techniques are applicable to complex sample designs in analytic studies.

Conclusions:

• Survey research methodologies provide essential tools for both evaluation and case-control studies.
• Adapting survey sampling techniques enhances the validity and efficiency of analytic study designs.
• Addressing challenges in control selection and employing advanced analysis methods improves the reliability of research findings.