453 



■"ii- 



vfaen vorklag hard" appearr in the same factor chese Cvo iteu have 

 opposice tignSy Indicating that those vfao smoke aost at the weekend tend 

 not to siBoke vdten voxking hard and conversely that those who smoke when 

 working hard tend »ot Co aooke aost at the weekend. 



It would, of coerse, have been possible to choose a solution with 

 a smaller number of factors, but only at the expense of losing some of 

 those factors to which McKennell attached particular importance. At 

 the 11-f actor level, for instance. Relaxation Smoking did not emerge 

 as a separate factor. As it stands, the 12-factor solution shows ^ood 

 agreement with HcKennell's factor structure, the replication being if 

 anything closer than between McKennell's first and second studies (4, 5). 



Table .3 shows the laatrix of intercorrelatiens among the 12 Promax 

 factors. The iirst five factors (1, 4, 12, S and 9); are those which 

 McKennell Included 3a his Inner Need dimension, and the aodest positive 

 correlations aacng then fit in with this concept. Three otner factors, 

 those of Refreshment Ssokiogj Idle Smoking and Continual Stnoking, are 

 also correlated with the 5 Iim©t Heed factors, and might perhaps be 

 regarded as forming part of the s^ae dimension. The factor which is 

 least correlated with all other factor* is Social Confidence. 



<^: 



03 



c 



I 



c 





