Too Old to Answer?: An Analysis of Randomized Trials of Mode of Survey Administration by Age

• Activity Number: 14
• Type: Contributed
• Date/Time: Sunday, August 11, 2002 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods*
• Author(s): Patricia Gallagher*+ and Lauren Harris-Kojetin and Vickie Stringfellow
• Affiliation(s): Center for Survey Research and RTI International and Center for Survey Research
• Keywords: Mode Effects ; Age Effects ; Telephone Survey ; Mail Survey ; CAHPS

Abstract

Researchers are concerned that cognitive declines associated with aging may compromise data collected from older pops. There is particular concern about tel surveys, where respondents are expected to remember all response choices until an answer is selected. Many surveys employ a multi-mode approach to data collection. To successfully combine data collected using multiple modes, mode effects must be minimized. We examine responses by mode in two age groups: 1.) 18-64 year olds (n=907); 2.) age 65+ (n=953). DESIGN: Cases randomized to mail or tel administration. MEAS: Consumer Assessment of Health Plans/CAHPS items common to both surveys. Four distinct response tasks tested. RESULTS: In both groups, there were bivariate differences by mode in eight of the fourteen items, although only four items were common to both age grps. Controlling for health status eliminated mode diffs in all but three items in the <65 grp, and all but four in the 65+ grp. Only one item demonstrated a mode effect in both grps. CONCLUSIONS: There is little consistent evidence that older people are any more likely than those younger to respond differently by mode to the items tested.