Volume 4: No. 3, July 2007
Triangulating Differential Nonresponse by Race in a Telephone Survey
This figure presents a flowchart on participant recruitment for baseline
and refusal interviews for the PRISM (Personally Relevant
Information about Screening Mammography) study. The sampling frame of eligible potential PRISM participants
included 27,944 women. Of these, E-Tech estimateed 26,688 names as black (n = 2448) or
white (n = 24,240). The random sample for PRISM telephone interviews
included 9079 women.
Of the completed baseline PRISM interviews
(n = 3543), 3490 (95.5%) self-reported race as black or
10.7% of 3490 self-reported as black (n = 374), and
89.3% of 3490 self-reported as white (n = 3116).
Excluded from the PRISM interviews were 5536 women. Of the 5536 women
excluded, 838 could not be contacted for the following reasons: 434 had no working telephone; 288 could
not be reached after the maximum number of 12 calls were made; 81 requested callback but could not reached; and 35 interviews were refused by gatekeepers. Two hundred and sixty women were deemed ineligible for the
following reasons: 125 had a mammography date outside range; 72 had breast cancer history/double mastectomy; 42 were too ill; 11 were not State Health Plan members;
there was a language barrier for 7 women; and 3 women were deceased.
In addition, 2422 women were removed from sample because target enrollment was met, and 2016 refused to participate. Of those who refused participation, 238 were invited to participate
in refusal interviews. Of these 238, 150 completed refusal interviews. Of these 150, the study team collected data on self-reported race and interviewer-estimated race
for 24 women. Of the 238 invited
to participate in refusal interviews, 88 declined to participate in refusal interviews, and of the 88 women, interviewers estimated race of 29 women.
Return to article