EXPENSE EACTOES EN" CITY DISTRIBUTION OF PERISHABLES 7 
METHOD OF DATA ANALYSIS 
The individual price records, received on printed cards, were tran- 
scribed to working sheets, and thence to punched cards for machine 
tabulation. 5 The information was thus classified according to the 
different conditions to be analyzed. In making comparisons of the 
different groups of data, the arithmetic mean of prices was used 
because this was found to be the most representative and the most 
workable kind of average. 
Since the purpose of the analysis is to compare spreads between 
wholesale and retail prices, under various market conditions, it is 
necessary to relate these prices to uniform bases. The method com- 
monly employed for this purpose is to express price spread in terms 
of a uniform outlay of $1 by the consumer. In this case the spread 
represents the number of cents of the consumer's dollar that are 
absorbed by the expense of city distribution. In other words, it is 
the difference between the value of the dollar's worth of goods at 
retail and the value of the same goods at wholesale. 
With such a base the spread is conveniently expressed also as a 
percentage margin. This may be derived by dividing the price differ- 
ence for any quantity of goods by the retail price of the same quan- 
tity. Thus, if R represents retail price, and W represents wholesale 
price for a given quantity of goods, the price difference, or spread, is 
represented by R-W. The percentage margin is then represented by 
E-W 
R 
A preliminary inspection of the average percentage margins in the 
various classifications of data was made to find out which of the 
different market factors considered were accompanied by the greatest 
margin differences. The widest divergence was found to exist be- 
tween the different individual commodities in the series of 14 articles. 
Next to the contrasts in commodities, the divergence of margins was 
greatest in different types of retail stores, as indicated by differences 
in management and in selling policy. These two sources of con- 
trast overshadowed all other distribution factors that were consid- 
ered. The analysis was therefore focused upon the factors which 
appeared to account for the greater amount of difference in distribu- 
tion expense: (1) The nature of the commodity and (2) the type of 
store operation. 
ANALYSIS OF COMMODITY DIFFERENCES 
The composite means of retail and wholesale prices of each com- 
modity as a whole for the entire period embraced in this analysis are 
shown in Table 3. Retail store prices are weighted according to the 
importance of different store types, and are adjusted for shrinkage in 
retailing. The average distribution expense per package is the differ- 
ence between mean retail price and mean wholesale price. It is 
expressed as a percentage of the retail price in the last column. 
These average prices and percentage margins are the basic figures 
used in making the commodity- analyses. 
s The work of tabulating and classifying the large number of compilations was done by the machine 
tabulation section of the Bureau of Agricultural Economics in Washington, under the direction of E. J. 
Way. 
