462 



-20- 



and tbac cIm WHEAT clusters cover ouch the same part of the graph as . 

 McKeonell's clusters. Points of deull are that, in the Social diaeosion, 

 Clusters 1 and 8 «r« respectively lover and higher than any of HcKenncll's 

 clusters; in tl»$ Inner Meed dunension Cluster 3 has the lowest score of 

 either set, while VfcKennell's CTuster VII has the highest score. 



The distribution of certain external variables over the clusters 

 is shown in Table 6. Inspection of the results indicates that some of 

 the external variables (average cigarette consvjsption, depth of inhalation, 

 difficulty in giving up) are related in a meaningful way with the level 

 of Inner Need. The three relationships are shown in Figures 2 to 4. 

 When it is remembered that the external variables of cigarette consumption, 

 depth of inhalation and difficulty in giving up jjmoking are quite independent 



of the data fcda which the Inner Need scores were d^^ived, the correlations 



, -y -- - 



depicted ara ratJi^ striking and support the content iou(jthat th« factors 

 on which the classification of smokers has been based real<y do reflect 

 motives for smoking. 



Based on the information^'^iven in Tables 3, S and 6 it is possible 

 to draw up the following detailed cluster profiles: 



Cluster 1: Low Need-Low Social . Ibis comprises the smallest group 

 (3.3Z).- The only departure from a unifor^y low profile on all factors 

 is the slight preference to smoke oiore at the yeekends. Cluster 1 can 

 be distinguished from the other Low Need clusters '^y the particularly 

 low score on the Social factor. Members of this cluster have the lowest 

 depth of inhalation and the second lowest cigarette consumption (14 « 



cigarettes per day) . Their average age is greater than that of the % 



o 

 total sample, but they are well distributed throughout the social class - 



o 



groups. "" 



