Random allocation to treatment occurred at the individual level within the questionnaire itself. The controlling variables were computed via system randomisation at the start of the individual interview. Each vignette required a unique set of six controlling variables, one for each life dimension, while the names were allocated such that only one control variable was needed. The controlling variables, to be assigned in the questionnaire and in the record i_indresp:
ff_vigname: sequence of names used in the three vignettes
1 George, Margaret, Harry
2 George, Margaret, Anne
3 George, Harry, Margaret
4 Margaret, George, Harry
5 Margaret, George, Anne
6 Margaret, Anne, George
ff_vdisease1 (for vignette1), ff_vdisease2 (for vignette2), ff_vdisease3 (for vignette3)
1 No long-term illness
2 Diabetes
ff_vdisability1 (vignette1), ff_vdisability2 (vignette2), ff_vdisability3 (vignette3),
1 No difficulties climbing stairs
2 Difficulties climbing stairs
ff_vcognit1 (vignette1), ff_vcognit2 (vignette2), ff_vcognit3 (vignette3),
1 No problems remembering
2 Problems remembering
ff_vphys1 (vignette1), ff_vphys2 (vignette2), ff_vphys3 (vignette3)
1 Opens food packaging easily
2 Struggles to open food packaging
ff_vinterp1 (vignette1), ff_vinterp2 (vignette2), ff_vinterp3 (vignette3),
1 Sees [his/her] family and friends regularly
2 Rarely sees [his/her] family and
ff_vvolun1 (vignette1), ff_vvolun2 (vignette2), ff_vvolun3 (vignette3)
1 Often volunteers
2 Doesn’t volunteer
The variables used for this experiment are in record i_indresp:
vign1, vign2, vign3



