Design and Operation of the 2010 National Survey of Residential Care Facilities. NONRESPONSE BIAS ANALYSIS


As previously noted, the NSRCF facility response rate weighted by the inverse of the probability of selection was 81%. Table C presents the percent distribution of selected facility characteristics among responding and nonresponding facilities. Using sampling frame variables, nonresponse bias was assessed by examining how much the respondents and nonrespondents differed on the survey variables of interest. Responding and nonresponding facilities were significantly different by size (small, medium, large, and very large) and geographic region (Northeast, Midwest, South, and West), but not by MSA status (metropolitan or nonmetropolitan).

Table C also presents the weighted and unweighted response rates among facilities with different characteristics. Higher cooperation was gained among small facilities and facilities in the Northeast and West. The effect of this differential response is minimized by using a nonresponse adjustment factor that takes into account size, region, and MSA status. These variables were chosen because they were deemed important for analysis and were also used as estimation strata. To evaluate the effect of nonresponse adjustments on selected survey estimates, the last column of Table C shows estimates weighted by the inverse of the probability of selection and adjusted for nonresponse. The adjustments reflected in the last column of Table C were applied to reduce bias in survey estimates.

View full report


"NSRCFdo.pdf" (pdf, 8.47Mb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®