461 



-19- 



o£ two McKttinell cluscert,. ▼ and VII, which tgaia can b« seen by cooparing 

 Che profiles. Four of Che WHEAT cluacera (2, 6, 5 and 10) have no close 

 HcKennell equivalent. 



An ioportai^ difference is Chac only 2SZ of che WHEAT sample was 

 classified as Low Seed at ooop'a'red wich 41Z of HcRennell's, Che biggesc 

 disparity being in Cluster 1. A possible explanation could be the 

 different age distribution* of the two samples, with S3Z of McKennell's 

 sample being 4S or older as against 37Z of the WHEAT sample. This could 

 affect the size of Cluster 1 which ia both cluster solutions was considerably 

 above the average age of the sample. -; 



McKennell laid emphasis on the distinction beCwecn the Inner Heed 

 component, (comprising information from the first five factors as already 



stated) and ~the Social component (containing information from the two 



'^A 'O 



factors Social Snibving and Social Confidence). He predated « diagram 



showing the dispersion of his clusters in these two dimensions. The 



precise way in which he treasured these two cooiponents was not followed 



/^ 

 in the WHEAT study. Bowever'^$Q approximation to his method of presentatioa 



can be arrived at by plotting the mean of the factor scores on the 



first five factors against the aean bf the factor scores on the two 



social factors for each of the 12 clusters. This plot is shown m 



Figure 1, which also gives the positions of "MpKennell's clusters after 



caking the means of the factor scores in the same way. Although much 



of the distinction between clusters is lost, this oteckod of presentation ^ 



c 

 does facilitate an overall comparison between the WHEAT clusters and c 



O 



McKennell's clusters in terms of these two dimensions. FiguT^' I shows O' 



Chat Che clusters in each sec are fairly well separated from one toother ° 



