A comparison of the estimated consumption functions for meat from two different samples

The objectives of this study were (1) to determine if a 1967 survey on meat consumption in Knoxville, Tennessee, conducted by Williams was taken from the same population as a 1965 survey on food consumption in urban areas of the South conducted by the Consumer and Food Research Division of the Agricultural Research Service, and (2) to compare the estimated consumption functions for meat for these two data sets.

The information presented in the two data sets was made as alike as possible. The two data sets were combined to form a third data set. These three data sets were analyzed using the same models. Ordinary least squares regression was utilized in the analysis.

Six models were fitted using both total quantity and total expenditures as the dependent variables. Models I and II assumed the dependent variables to be functions of annual household disposable income and 17 other socio-economic factors while Models III, IV, V, and VI assumed total quantity and total expenditures were functions of income only. Income for Models I and III were coded from 1 to 14 representing 14 different income groups. For Models II and IV, income was treated as 14 different 0,1 dummy variables, a dummy variable representing each income group. These income groups were as follows: under $1,000; $1,000-1,999, $2,000-2,999; $3,000-3,999, $4,000-4,999; $5,000-5,999; $6,000-6,999; $7,000-7,999, $8,000-8,999, $9,000-9,999; $10,000-11,999; $12,000-14,999; $15,000-24,999; and $25,000 or over. Model V measured income in groups of $0-2,999; $3,000-5,999; $6,000-8,999; $9,000-11,999; and $12,000 or over. Model VI measured income in groups of $0-1,999; $2,000-3,999; $4,000-5,999; $6,000-7,999; $8,000-9,999; $10,000-14,999; and $15,000 or over. There existed a 0,1 dummy variable for each income group for Models V and VI.

Varying results were obtained from the use of these six models. Different amounts of variation in total quantity and total expenditures were explained for the Knoxville Survey and the '65 Household Survey. These models also found different variables significant for the two data sets. The same sign did not always appear for a given variable for both data sets. However, a general dip in consumption in both data sets did occur.

By using a form of the F test, it was concluded that the Knoxville Survey data and the '65 Household Survey data did not have the same relationship to meat consumption and expenditures for meat.