472 



-y>- 



16 per week for chose la Cluacer 3 Co 51 per week for those In Cluster 12. 

 So it CAii be said chat t proportion of respoodeots suppleoent their 

 snokiag of filter-tipped cigarettes with other cigarettes of higher 

 oicotinc delivery, tb« proportion who do so and the level of coosuapcLoa 

 increasing along wltb. Inner lecd. Cigar and pipe smoking is also indulged 

 in hj a .fair nuaber of '^lokera,. but consumption levels are low averaging 

 2-3 cigars per week and on^ ounce of pipe tobacco per week and not varying 

 appreciably between clusters, ^e incidence of cigar and pipe snoking 

 is related more to age and social ^lass than it is to cluster membership. 

 A^^, Concern for Health ^''y 



Aa aientioned in the section on Methods^ that part of the questionnaire 

 relating t< concern for the possible ill-«f fec(4 pf ssuking was not 

 asked until respondents had smoked the various expe(^ent«l cigarettes 

 and had been questioned about them in the final recall^v^^ervicv. It 

 was felt that any earU^ exposure to the health-related questions ai^t 

 bias their opiniona of thif experinental cigarettes. Naturally some 

 fall-out of respondcn?* oecurr^ between the initial survey and the final 



recall interview, added to which %ft«^%fhole of Cluster 2 was deliberately 



'^. 

 eliminated from the cigarette Ccating'^rt of the exercise for reasons 



which will be explained in the subsaquent Veport (i.e. Tart 2). Tb« . 



result was that the health-related portion of^ihe questionnaire was 



administered to a considerably reduced sample of respondents (1023, 



compared with 1523 in the initial survey) and there iay have been soae ^ 



/ c 



shift in cluster profiles because of the fall-out. • c 



The answers to 7 of the questions were scored, as explainad at the o 



ff> 



end of Appendix V, to give an overall score for health concern for each '^ 



