LIFE HISTORIES AND HEALTH
IN MID-LIFE STUDY:
FINAL RESPONSE RATES
LIFE HISTORIES AND HEALTH STUDY
The Life Histories and Health Study utilized a sub-sample of subjects from the Wisconsin Longitudinal Study (WLS) to explore relationships between cumulative life experience and health (physical and mental) among men and women in their late 50's. This document provides an overview of the recruitment process, summarizes the results of that process, and finally, provides some interpretation of those results. Our goal was to collect survey data (via home interviews and mail questionnaires) from approximately 200 individuals and biological data from about 80 persons in that group.
Subject recruitment occurred in three phases beginning in October 1996. In the first two phases a letter, from Bob Hauser, was sent to potential subjects. The letters included a brochure about the Life Histories Study and a form subjects could use to either request more information about the study or decline the opportunity. Individuals who did not respond to this letter within 3-4 weeks received a second letter, and then a third if necessary. A small group of individuals did not respond to any of these letters. Thus, as the final step in this part of the recruitment process, Debbie Carr contacted these individuals by phone to determine if they wanted additional information about the study. If a subject requested more information their name was forward to the Life Histories Study, a written invitation to participate was extended, and finally subjects were contacted by phone so that questions could be answered and participation confirmed. The third recruitment phase was targeted at individuals who originally consented to participate in just the survey portion of the study.
All subjects were drawn from sampling group 0 (n=113) and 1 (n=131) of the larger WLS. They were selected a priori according to criteria outlined in the May 22, 1996 memo from Ryff et al to Bob and Tess Hauser (attached).
10/1996 - 4/1997
Letters were sent to 244 individuals inviting them to participate in both the survey and biological components of the study.
Approximately 130 individuals had agreed to participate in the survey portion of the study, about 70 of these individuals had also agreed to participate in the biological component.
7/1997 - 8/1997
The selection criteria were expanded, adding 110 individuals from sampling groups 0 and 1 to the recruitment pool.
Letters were sent to these 110 individuals inviting them to participate only in the survey component of the study.
Survey data collection completed, the final sample size is 202 (112 women, 90 men).
Preliminary analyses of biological and survey data reveal that the biological sample size was too small to make meaningful inferences about relationships between cumulative life experience and present physical health.
Individuals recruited in the 9/1997 for survey data collection were invited to also participate in biological data collection. Some individuals in our original recruitment pool indicated that were not able to participate in the biological data collection at that time, but might be able at a later data. Thus a second invitation to participate in biological data collection was also extended to these individuals.
Biological data collection was completed, 122 individuals completed all or part of the biological component of the study (men n=61, women n=61)
Overall response rates and detailed summaries of response rates for the full sample (and by gender) are provided below. These are further broken down by the different components of participation (i.e. interview, mail questionnaire, GCRC/Davidson lab, immune).
Overall Response Rates
Response Summary for the sample (n=354), selected a priori from WLS sampling group 0 and sampling group 1.
Whole Sample Men Women
(N=354) (n=158) (n=196)
Deceased 4 (1.1%) 3 (1.9%) 1 (0.5%)
No Address 6 (1.7%) 4 (2.5%) 2 (1.0%)
No Response 26 (7.3%) 10 (6.3%) 16 (8.2%)
No to Letter/call from 78 (22.0%) 34 (21.5%) 44 (22.4%)
Yes to Letter/call from 240 (67.8%) 107 (67.7%) 133 (67.8%)
Response Rates for Different Components of Participation
Among the 240 individuals who said "Yes, I'd like more information", 202 (84%) completed the survey portion of the study (interview & mail questionnaire); 90 men (44.5%) and 112 women (55.4%)*. Within that group, 122/202 (60.4%) also participated in at least one component of the biological data collection (GCRC, Immune); 61 men (67.7%) and 61 women (54.5%). A more detailed summary of the proportion of respondents completing each part of the study is provided below.
*There were 38 additional individuals who requested additional information, but ultimately declined participation (see below for more information).
Whole Sample Men Women
(N=240) (n=107) (n=133)
Declined Participation 38 (15.8%) 17 (15.8%) 21 (15.8%) (after receiving more information)
Interview/Mail only 80 (33.3%) 29 (27.1%) 51 (38.3%)
Interview/Mail/GCRC 21 (8.7%) 8 (7.5%) 13 (9.7%)
Interview/Mail/Immune 16 (6.7%) 4 (3.8%) 12 (9.0%)
Immune 85 (35.4%) 49 (45.8%) 36 (27.1%)
Reasons for Declining Participation
The 38 individuals who declined participation after getting additional information fall into one of 2 groups. The first group (27/38) includes those persons who declined participation after learning more about the study. The second group (11/38) declined at some point in the process of scheduling or attempting to complete the interview. The reasons for declining are summarized below (note some respondents gave multiple reasons for declining):
No time, too busy 10
respondent health 1
spouse/family health 2
family problems 2
too personal 1
just retired/moving out of state 1
Scheduling Issues could never reach R to schedule interview 2
after repeated attempts to schedule R declined or interview period ended 4
respondent missed interview, never rescheduled 1
interviewer couldn't find R's home, R declined 2
Issues Specific to Study
Not interested 3
Study too involved 3
Respondent declined after receiving mail questionnaire, Q's inappropriate 2
No Reason Given 7
Reasons for Declining GCRC/Davidson lab and/or Immune Component
Among the 202 individuals who agreed to participate in the study 39.6% completed the interview and mail questionnaire, but declined participation in the biological components of the study. Among the remaining 122 individuals, most (69.6%) completed the full biological assessment (GCRC/Davidson lab & Immune). The reasons respondents declined one or both of the biological components are summarized below.
Work Related Issues
works 2 jobs 1
works 6 days week 1
has dairy farm, can't get away 1
work schedule 4
doesn't want to come to Madison 5
too far 4
inconvenient/time consuming, 3
don't like city 1
no transportation 2
was in accident can't drive for 6 months 1
life is too overwhelming at this time 1
family/friend health 3
respondent's health 8
no time after all 6
doesn't need physical 1
struggling w/emotions since interview, not comfortable going to Madison 2
too personal 1
don't want to see strange doctor 1
Not interested 4
No Reason Given 8
Don't want vaccinations 1
Doesn't like needles 1
Knows people who reacted badly to flu shot - don't want to take that chance 3
Respondent got shot before we could give 2
R allergic/reacts to flu vaccine 6
R sick at time of vaccination, declined. 1
Blood hard to draw, respondent declined 3
No Reason Given 2
Does Response Rate and Degree of Participation Vary by Distance From Madison?
For the interview/mail questionnaire and immune components of the study one our staff traveled to the respondents home. Subjects completing the GCRC/Davidson lab components, however, had to travel to Madison. Thus we raise the question "do response rates and degree of participation vary according to distance from Madison?" If the answer is 'Yes' we would expect to see declining response and participation rates as the distance from Madison increases. We explore this question via the following summary tables. The distances categories were defined a priori by the WLS staff.
Table 1 (below) reports the distribution of subjects across the 6 geographic areas. Approximately 70% of the respondents live within 120 miles of Madison. The largest cluster of subjects live within the 40-80 mile range and the smallest within the 160-200 range. Men and women are distributed in similar patterns across areas with one notable exception; there are almost twice as many women, compared to men, living 200+ miles from Madison.
Table 1: Distribution of sample across geographic areas for the whole samples, men and women.
|Distance from Madison||Distribution Across Geographic Areas|
|Whole Sample (N=351)*||Men
|0 - 40 miles||50 (14.2%)||19 (12.2%)||31 (15.8%)|
|41 - 80 miles||127 (36.2%)||56 (36.1%)||71 (36.2%)|
|81 - 120 miles||70 (19.9%)||29 (18.7%)||41 (20.9%)|
|121 - 160 miles||41 (11.7%)||23 (14.8%)||18 (9.2%)|
|161 - 200 miles||21 (5.9%)||15 (9.7%)||6 (3.1%)|
|200 + miles||42 (11.9%)||13 (8.4%)||29 (14.8%)|
* Distance data is unavailable for 3 of the 354 cases, all men.
Response rates for the whole sample do not appear to vary in a patterned way across geographic areas. However, as Table 2 (below) suggests, willingness to travel to Madison may be influenced by distance to Madison. Among persons living in the 0-120 range, rate of participation in the GCRC/Davidson lab component decreases as distance to Madison increases. A similar pattern is observed among those living more than 120 miles from Madison. The highest participation rate among all areas is observed for persons living in the 121-160 mile range with rates decreasing, as distance increases, from there.
Table 2: Response and participation (GCRC vs. no GCRC) rates for the whole sample.
|Distance from Madison||Whole Sample|
|Willing to Participate||Willing to come to Campus|
|0 - 40 miles||29/50 (58.0%)||18/29 (62.1%)|
|41 - 80 miles||76/127 (59.8%)||38/76 (50.0%)|
|81 - 120 miles||36/70 (51.4%)||15/36 (41.6%)|
|121 - 160 miles||24/41 (58.5%)||18/24 (75.0%)|
|161 - 200 miles||13/21 (61.9%)||7/13 (53.8%)|
|200 + miles||24/42 (52.1%)||10/24 (41.6%)|
The final table (Table 3) in this series reports response and participation rates across geographic areas for men and women. The Response rates vary widely across areas for both men and women, but with no discernible pattern. The gender specific distribution of participation rates, however, provides additional insight regarding the pattern noted for the whole sample above. Distance from Madison seems to be more of an issue for women, compared to men, living in closest 3 areas (0-120 miles); rates of participation for women decrease as distance increases. Surprisingly, the opposite is true for persons living more than 120 miles from Madison. There is a steady decrease in participation rates among men as the distance to Madison increases, but not women. Thus we suggest that distance is linked to response rate, although there appears to be gender-linked variability in participation rates as distance increases.
Table 3. Response and participation (GCRC vs. no GCRC) rates for men and women, separately.
|Distance from Madison||Willing to Participate||Willing to Come to Campus|
|0 - 40 miles||13/19 (68.4%)||16/31
|41 - 80 miles||29/56 (51.8%)||47/71
|81 - 120 miles||13/29 (44.8%)||23/41
|121 - 160 miles||16/23 (69.6%)||8/18
|161 - 200 miles||11/15 (73.3%)||2/6
|200 + miles||8/13