VIII 
INTRODUCTION 
THE discussion in Part II was concerned primarily with tracing 
out norms and trends in the individual series of data relating 
to member banks in the different Federal Reserve districts, three 
bases of comparison for the most part being used. These were 
as follows: first, the nature and the percentage amounts by 
which the respective ratios for member banks in the several 
districts differed from their own seven-year average levels; sec- 
ond, the directions and the percentage amounts of change in 
the ratios from year to year; and third, the nature and per- 
centage amounts by which the ratios in the respective districts 
each year deviated from the average ratio for the country as 
a whole. These three sets of comparisons were made for each 
of the types of data available, the norms and trends being 
determined by summarizing in various ways the experiences in 
the different districts, by observing the uniformities. and by 
generalizing as to their preponderance. 
In Part III the purpose of the discussion is different, although 
the method of attack and the bases of comparison are much 
the same as those just described. Interest now is not in the 
discovery of norms and the measurement of trends in individual 
series of data, compounded out of the experience of the banks 
in the several districts year by year; it is rather in observing 
and measuring the interrelations of norms and trends in fwo 
or more series at the same time. 
Previous analysis has shown that while variations characterize 
the ratios in the individual series, they are not generally of the 
random sort. Are there also uniformities and tendencies com- 
mon to the paired variations and changes in different series? 
Moreover, in what way and to what degree are the series re- 
lated? Specifically, such queries as the following arise: In what 
manner are variable ratios of gross earnings, of operating 
expense, and of net earnings associated with variable ratios of 
" 
12