Sec. 7.5 



RANK CORRELATION 



229 



were obtained from the Department of Home Economics, Kansas Agricultural 

 Experiment Station, through the courtesy of Dr. Abby Marlatt. 



Height in Centimeters (X) 



Mean 



weight 246.6 259.9 285.4 306.8 336.5 378.7 409.9 458.1 



Standard 

 deviation 29.1 27.2 35.6 38.6 57.9 109.0 82.8 



11. Plot the mean weights above against the midpoints of the height classes, 

 and decide therefrom if the assumption of a linear relationship between these 

 two variables seems acceptable. 



12. Ignore any indication of non-linearity of trend and compute r and b by 

 the methods of this chapter. What conclusions can you draw from these 

 estimates? 



Arts, b = 0.77, r = .72. 



13. Compute the standard deviation not given in the above table by the 

 method of Chapter 2 adjusted so as to take account of the fact that this is 

 supposed to be a sample. 



14. Each height class has some kind of one-variable frequency distribution 

 of the weights within the height class. Hence the above data constitute 

 several samples of weights within height classes. Theoretically, these weight 

 distributions within height classes must have equal population variances. Plot 

 the standard deviations against the midpoints of the corresponding height 

 classes and decide therefrom — if you can — whether or not that is a good as- 

 sumption in this case. 



15. For the weight class 290 to 329 kilograms compute the coefficient of varia- 

 tion for the heights, taking the point of view maintained in Chapter 2. 



