ESRA 2027 sessions by theme

Back to Overview of Sessions

Session Details

In survey research, we are often interested in the reliability and validity of the measures we use to evaluate the quality of measurements we can obtain. Such a focus typically relies on the individual-level correlations between attitude measures (split-half measurement; test-retest reliability measurement) or between attitude measures and behaviors (concurrent or predictive validity). These individual measures are often developed using the psychometric tradition which designs for testing and evaluations of individuals, not of large samples. As a result of psychometric influences, survey researchers often choose scales to use that have many long, and often, very repetitive items to include in sample surveys. However, in the sociometric approach (Thomas, 2016), sufficiently large samples can attain high levels of stability and validity of measurement at the group level, meaning the results from a single item can be sufficiently stable and robust to detect difference across groups and across time thus obviating the need to use many repetitive items for research purposes (consequently reducing respondent burden). Our interest is in this session is to focus on developing group-level measures for both reliability and validity, some of which may involve repeated sampling to establish measurement parameters of reliability and validity. We invite a variety of submissions that can advance this goal by developing techniques or measures that can be useful to assess either the reliability or validity of data from larger samples (e.g., over 100 cases). The goal of this is to develop a sample-level approach to measurement parameters distinct from individual-level derived from the psychometric tradition used for individual assessment. A focus will be on the utility of such measures for documenting both the stability and validity of sample survey assessment for description and inference.