Thus, an average 1950 expenditure of wage-earner and clericalworker families, adjusted for sampling and reporting errors, and the abnormal spending pattern following the Korean outbreak, was estimated for each item of expense; and these item expenditures were related to average 1950 family income.
The second step in creating weights for the revised index adjusted these average expenditures to the fiscal year 1951-52. This was accomplished by applying the stable income-expenditure relationships established for 1950 to estimated average family incomes for the 1951-52 period. This fiscal period was taken only because necessary information for the calendar year 1952 was not available at the time. Average 1951-52 incomes were estimated from regressions of 1950 average family income from the Consumer Expenditure Survey on average gross weekly earnings of production workers.
Since various local factors other than income affect the level of housing expenditures and relationships among component parts of total housing costs, no income-expenditure relationship could be established for this group. The ratio of homeowners to renters had not changed significantly between 1950 and the 1951-52 period. Therefore, the expenditure weights for housing required only the adjustment of the 1950 reported expenditures on rent, fuels, repair, and maintenance items, for price changes since 1950. For homeownership costs, estimation of expenditures for home purchase and interest payments and the rate of home purchase required to maintain the 1951-52 level of homeownership were developed from the 1950 survey and from census data.
The index "market basket" thus represents the customary buying pattern of city wage-earner and clerical-worker families in the period 195152. It includes television receivers and frozen foods that were not a part of family living patterns a few years ago, and it includes other important changes that have occurred in the amounts, kinds, and qualities of things people buy.
that are relatively important in family spending, items that are representative of price change for large groups of related commodities, and items that have distinctive price movements of their own. In some cases, several qualities are priced to represent a single item.
During 1950 and 1951 the Bureau priced and studied the price changes of hundreds of items in order to expand and bring up to date the price information available in the Bureau's records. Items were then stratitied within groups having similar characteristics with respect to physical description, use, and other price determining characteristics, into "price families" of items whose prices fluctuate similarly over reasonably long periods of time. Within each "price family" those items which were of outstanding importance in family spending were selected to represent price changes on all other items in the group. The number of items selected depended on the prevalence of items within a "price family" with high relative importance to total expenditures. %%There only one item Was selected, it was assigned the total expenditure weight of the group it represented; where two or more items were chosen, the total weight of unpriced items was assigned to them proportionate to their own importance in the group. Since price relationships change over time, the "price families" of items established for index item selection are reexamined periodically to determine whether price-change imputation patterns require adjustment.
Among the 300 items included in the sample are all the goods and services that arc outstanding in family purchases, so that the priced items directly represent the greater part of total family expenditures. Intercity differences in the list of priced items were limited to commodity groups where differences in the kinds of goods purchased were highly .
such as fuels and certain types of clothing, in order to standardize the pricing procedures as much as possible. Small differences in qualities of items offered for sale in the respective cities arc reflected in city-to-city variations in the descriptive specifications used in obtaining current prices.