Sas software program for windows

All analyses were conducted using Version 9.4 of the SAS® software program for Windows (copyright © 2013, SAS Institute Inc., Cary, NC). First, an item-level data set was created that contained the item characteristic values and acquiescence variables for each of the 90 heterogeneous items. As the data set was organized by items instead of respondents, respondent characteristics could not be added to this data set. In order to examine acquiescence for particular respondent characteristics, additional item-level acquiescence variables were created using data drawn only from respondents with selected respondent characteristics. Stratified analyses were then conducted within these respondent subgroups to examine acquiescence and item characteristics by respondent ethnicity (four subgroups), education (high school or less vs. more than high school), age (18-50 vs. 51-90 years), interview language (English vs. Spanish), and gender (male vs. female). The limited cell sizes prohibited simultaneous examination of more than one respondent characteristic (e.g., estimating acquiescence for 18-50 year-old Mexican Americans). All hypotheses were tested through general linear models regressing acquiescence on multiple item characteristics. Since acquiescence was assessed at the item and not the respondent level, no control variables were included.