Correlates of Consumption of Food Nutrients By U.S. Urban Households

Mohamed Abdel-Ghany, University of North Carolina-Greensboro
Gordon E. Bivens, University of Missouri-Columbia
ABSTRACT - Nutrient and calorie intakes of U.S. urban households were examined in relation to six economic and socio-demographic characteristics. Adjusted Household Meal Units was most frequently, and positively, associated with nutrient intakes. Income was positively associated with five individual nutrients. Education and age of female homemakers tended to be negatively associated with nutrient and calorie intakes.
[ to cite ]:
Mohamed Abdel-Ghany and Gordon E. Bivens (1976) ,"Correlates of Consumption of Food Nutrients By U.S. Urban Households", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 229-237.

Advances in Consumer Research Volume 3, 1976      Pages 229-237

CORRELATES OF CONSUMPTION OF FOOD NUTRIENTS BY U.S. URBAN HOUSEHOLDS

Mohamed Abdel-Ghany, University of North Carolina-Greensboro

Gordon E. Bivens, University of Missouri-Columbia

ABSTRACT -

Nutrient and calorie intakes of U.S. urban households were examined in relation to six economic and socio-demographic characteristics. Adjusted Household Meal Units was most frequently, and positively, associated with nutrient intakes. Income was positively associated with five individual nutrients. Education and age of female homemakers tended to be negatively associated with nutrient and calorie intakes.

In recent years much attention has focused on the deterioration of the nutritional level of American diets. Despite the availability of sizeable and diverse food supply, some Americans consume less than recommended dietary allowances of food nutrients. Unequal distribution of economic resources in America has led to the existence of pockets of poverty and associated malnutrition; however, some high income households also suffer from malnutrition either because of lack of nutritional information or because of affluent life styles. This paper examines how households of differing socio-economic characteristics vary in their consumption of food nutrients.

METHOD

Sample

The source of data for this study was the U.S.D.A. Household Food Consumption Survey of 1965-66, the most recent nationwide survey of household food consumption. The sample was limited to the data collected in the spring of 1965 (the largest single wave of four seasons of interviewing); also the sample was confined to the urban black and white households. Households with missing responses for any of the variables used in the analysis were eliminated. Accordingly, 3,860 households constituted the studied sample.

The Dependent Variables

Designated as dependent variables were per household weekly actual consumption of food energy and nine nutrients -- protein, fat, calcium, iron, vitamin A, thiamin, riboflavin, niacin, and ascorbic acid.

The Independent Variables

The independent variables included: the education of female homemaker, household income after taxes, age of female homemaker, race of female homemaker, region of the country (Northeast, North Central, West, South), and Adjusted Household Meal Unit. The measure of Adjusted Household Meal Unit (AHMU) was devised to take into account differences among households in number of meals eaten at home by the various members as well as the household size and composition. The procedure for the derivation of the AHMU variable was as follows:

(1) consumption equivalence scales for 20 sex-age categories were computed based on the USDA moderate cost food plan (U.S. Department of Agriculture, 1965). The 20-35 year old male category is used as the base; the cost of one week's food for each of the other sex-age categories is divided by the estimated cost of one week's food for a 20-35 year old male. The consumption equivalence coefficients for the 20 sex-age categories are presented in Table 1. (2) The total number of meals served each member of the household per week is multiplied by the consumption equivalence coefficient applicable for each household member's sex-age category. (3) The sum of the adjusted meals for all members of the household is calculated. The AHMU, then, is

EQUATION

(the total meals served per week to each household member x consumption equivalence coefficient applicable for each household member)

where i = the number of household members

Method of Analysis

Step-wise least squares multiple regression was used as the technique of analysis since it permits examination of the effects of each independent variable on a given dependent variable while controlling for the simultaneous effect of other independent variables. This technique is capable of handling variables measured in either quantitative or qualitative terms. Within the regression models, qualitative variables appear as sets of dummy variables; these dummy variables are given a value of one if an observation falls in a designated category, zero if it does not.

TABLE 1

COST OF ONE WEEK'S FOOD UNDER THE U.S.D.A. MODERATE-COST FOOD PLAN AND CONSUMPTION EQUIVALENCE SCALE COEFFICIENTS FOR DIFFERENT SEX-AGE CATEGORIES, 1965

The Statistical Model

The multiple regression model used to test the significance of the different household characteristics may be represented as follows:

Y = a + b1X1 + b2X2 + b3X3 + b4X41 ..... + b6X43 + b7X5 + b8X61 ..... + b10X63

where

Y is the dependent variable,

a is a constant term,

X1 is the household income after taxes,

X2 is the adjusted household meal unit,

X3 is the education of female homemaker,

X41 ... X43 is an expression for age of the female homemaker in the form of dummy variables,

X5 represent the race of the female homemaker in the form of a dummy variable,

X61 ..... X63 is the region variable in the form of dummy variables,

and

b1 ..... b10 are the regression coefficients.

It should be noted that in testing for the significance of a dummy variable coefficient, we are testing to see if that coefficient is significantly different from that of the omitted category. For continuous variables, however, the coefficients indicate the direction and magnitude of the effect on the dependent variable of a one unit increase in the continuous independent variable.

TABLE 2

SUMMARY OF ASSOCIATIONS BETWEEN HOUSEHOLD CHARACTERISTICS AND NUTRIENT CONSUMPTION. ('+' INDICATES POSITIVE ASSOCIATION STATISTICALLY SIGNIFICANT AT .05; '-' INDICATES NEGATIVE ASSOCIATION STATISTICALLY SIGNIFICANT AT .05)

RESULTS

A summary of results is presented in Table 2 and details for each nutrient are reported in Tables 3 through 12. Overall, the percentage of variation in the household consumption of individual nutrients which was explained by the six independent variables was: energy, 61; protein, 59; fat, 55; calcium, 55; iron, 55; vitamin A, 16; thiamin, 58; riboflavin, 54; niacin, 53; ascorbic acid, 31. In the further discussion, it should be noted that the significance of the regression coefficients is measured by the F statistics. [Since there is one degree of freedom to be associated with the explained sum of squares, then F = t2. The test of significance in this case using F or t is identical. For further discussion see: (Snedecor and Cochran, 1956).]

Household income after taxes was introduced in the multiple regression model as a continuous variable. [The ordinally-scaled income variable in the data was treated as an interval variable in the analysis. Labovitz (1967) argues that the results obtained using ordinal data is rarely different from the results obtained using interval data.] Income had a statistically significant effect on the consumption of all nutrients with the exception of energy and the nutrients calcium, iron, thiamin, and riboflavin. The positive effect of income was statistically significant at the 0.001 level for fat, niacin, and ascorbic acid; it was statistically significant at the 0.01 level for protein, and at the 0.05 level for vitamin A.

One factor which may have dampened the expected positive relationship between income and the consumption of some of the individual nutrients might be a negative relationship between purchasing efficiency of nutrients and income. Abdel-Ghany (1974) indicated that the quantities of nutrients that were obtained per dollar decreased as household income increased. In other words, at higher income levels, it would appear households substitute more expensive and less nutritious foods for cheaper and more nutritious foods. Kellerman (1972) also found no significant relationship between income and the daily recommended food serving adequacy.

AHMU was expected to be positively related to the consumption of individual nutrients, since its formulation depends mainly upon family size, composition, and number of meals eaten at home. Clearly, that was the case; the coefficients of AHMU were statistically significant at 0.001 level for all nutrients, and the relationship was positive for all nutrients.

TABLE 3

ESTIMATED REGRESSION COEFFICIENTS AND F VALUES FOR SELECTED HOUSEHOLD CHARACTERISTICS AND CONSUMPTION OF ENERGY (IN KILOCALORIES)

TABLE 4

ESTIMATED REGRESSION COEFFICIENTS AND F VALUES FOR SELECTED HOUSEHOLD CHARACTERISTICS AND CONSUMPTION OF PROTEIN (IN GRAMS)

TABLE 5

ESTIMATED REGRESSION COEFFICIENTS AND F VALUES FOR SELECTED HOUSEHOLD CHARACTERISTICS AND CONSUMPTION OF FAT (IN GRAMS)

TABLE 6

ESTIMATED REGRESSION COEFFICIENTS AND F VALUES FOR SELECTED HOUSEHOLD CHARACTERISTICS AND CONSUMPTION OF CALCIUM (IN MILLIGRAMS)

TABLE 7

ESTIMATED REGRESSION COEFFICIENTS AND F VALUES FOR SELECTED HOUSEHOLD CHARACTERISTICS FOR CONSUMPTION OF IRON (IN MILLIGRAMS)

TABLE 8

ESTIMATED REGRESSION COEFFICIENTS AND F VALUES FOR SELECTED HOUSEHOLD CHARACTERISTICS AND CONSUMPTION OF VITAMIN A (IN IU)

TABLE 9

ESTIMATED REGRESSION COEFFICIENTS AND F VALUES FOR SELECTED HOUSEHOLD CHARACTERISTICS AND CONSUMPTION OF THIAMIN (IN MILLIGRAMS)

TABLE 10

ESTIMATED REGRESSION COEFFICIENTS AND F VALUES FOR SELECTED HOUSEHOLD CHARACTERISTICS AND CONSUMPTION OF RIBOFLAVIN (IN MILLIGRAMS)

TABLE 11

ESTIMATED REGRESSION COEFFICIENTS AND F VALUES FOR SELECTED HOUSEHOLD CHARACTERISTICS AND CONSUMPTION OF RIBOFLAVIN (IN MILLIGRAMS)

TABLE 12

ESTIMATED REGRESSION COEFFICIENTS AND F VALUES FOR SELECTED HOUSEHOLD CHARACTERISTICS AND CONSUMPTION OF ASCORBIC ACID (IN MILLIGRAMS)

Education of female homemaker had a negative effect on the consumption of all individual nutrients with the exception of ascorbic acid. The negative effect was statistically significant at the 0.001 level for protein fat, iron, thiamin, and niacin as well as for energy; it was statistically significant at the 0.05 level for riboflavin. The positive effect for ascorbic acid was statistically significant at the 0.05 level.

The findings of this paper with respect to the negative relationship between the education of the female home- maker and the consumption of most of the individual nutrients are congruent with the findings of Madden and Yoder (1972) as well as Kellerman (1972) who found that education was not significantly related to adequacy of nutritive intakes and the daily recommended food servings adequacy in their respective studies. Also, Wilson and Lamb (1968) conducted a study of food beliefs among women and concluded that college graduates composed the largest group who accepted questionable beliefs about food. This negative relationship is perhaps due to the higher opportunity cost of time in the home among more highly educated women. The higher the value of the homemaker's time based on her earning opportunities outside the home, the higher the cost of preparing meals at home, and the fewer meals prepared. This might lead to either simple preparation of meals at home, including using foods that come into the home in highly semi-prepared form, or eating away from home more often. In either case, the attainment of a well balanced diet might be secondary to the minimization of time spent in food preparation.

The partial regression coefficients for the age of female homemaker set of dummy variables indicate that the 65-and-older category showed a statistically significant negative difference from the omitted category at the 0.001 level with respect to all nutrients with the exception of vitamin A and ascorbic acid. The 45-64 years of age category showed a statistically significant negative difference from the omitted group at the 0.001 level for energy and calcium; at the 0.01 level for protein, fat, and riboflavin; at the 0.05 level for iron, thiamin, and ascorbic acid. The 14-24 years of age group showed a statistically significant negative difference from the omitted category at the 0.01 level only for niacin and ascorbic acid.

Young, Waldner, and Berresford (1956) found a negative relationship between age and nutritional level. Jalso, Burns, and Rivers (1965) concluded that as age increased, valid nutritional opinions and practices decreased. Madden and Yoder (1972) found that age of the homemaker has a negative effect on the adequacy of all nutrients with the exception of iron. Kellerman (1972) also found a significant negative relationship between the age of homemaker and the daily recommended food serving adequacy

Black households consumed less calcium and niacin (significant at the 0.001 level) as well as riboflavin (significant at the 0.01 level) than did white households. Black households consumed more vitamin A than did white households (significant at the 0.001 level). There was no significant difference in the consumption of energy, protein, fat, iron, thiamin, and ascorbic acid between black and white households.

Regional variations in food consumption and nutrition may arise because of differences in customs, costs of living, availability of certain foods, and effects of climate. However, in this study relatively few differences were found between household characteristics and household consumption of individual nutrients.

The partial regression coefficients of the Northeast region were negative and statistically significantly different from the omitted region (i.e., the South) at the 0.01 level for energy and fat; positive and statistically significantly different at the 0.001 level for vitamin A and ascorbic acid; and at the 0.05 level for riboflavin.

North Central showed to be positive and statistically significantly different from the omitted region at the 0.05 level for only niacin.

The Western region was negatively and statistically significantly different at the 0.05 level for fat; positively and statistically significantly different at the 0.01 level for vitamin A.

SOME IMPLICATIONS

Clearly, there is need for better understanding of nutrition and of the need for food consumption patterns which better fulfill nutritional needs. This prevails regardless of education or income and, for the most part, age. Therefore, it would appear there is a very present need for education, broadly perceived, both formal and particularly informal (e.g., extension programs, etc.) aimed at adults as well as younger people and probably in various educational modes. As part of this, possibilities may exist for imaginative use of short spots dealing with nutrition concepts for TV or radio airing. Other communication techniques undoubtedly have pertinence, too.

Perhaps the food processing industry needs to continue to examine its role in supplying the food and nutritional needs of the population. It may be that fortification, always in consultation with the best nutritional advice, needs to be re-examined, modified and carried further in some instances; particularly in foods commonly consumed by older people may this need exist. It may be especially important for the food industry to examine its advertising practices and to adopt a code aimed at providing information really useful to consumers in making choices affecting their nutritional status.

In any event, barring substantial shifts in the underlying relationships, it would appear that increases in general educational levels or an upward drift in incomes can't be assumed to bring an automatic improvement in household nutritional status.

REFERENCES

Abdel-Ghany, M. "Influences of U. S. Urban Household Characteristics on Nutritive Intakes and Quality of Food Consumption," Ph.D. dissertation, University of Missouri, Columbia, 1974.

Jalso, S. B., M. M. Burns, and J. M. Rivers. "Nutritional Beliefs and Practices," Journal of the American Dietetic Association, 47 (October 1965), 263-68.

Kellerman, B. J. "Correlates of Food Shopping Efficiency, Dietary Adequacy, Food Variety, and Food Expenditures Among Low-Income Rural Missouri Households," Ph.D. dissertation, University of Missouri, Columbia, 1972.

Labovitz, S, "Some Observations on Measurement and Statistics," Social Forces, 46 (December 1967), 151-160.

Madden, J. P. and Marion Yoder. Program Evaluation: Food Stamps and Commodity Distribution in Rural Areas of Central Pennsylvania. Agricultural Experiment Station Research Bulletin 780. University Park, Pa.: The Pennsylvania State University, 1972.

Snedecor, G. W. and William Cochran. Statistical Methods. 5th ed. Ames, Iowa: The Iowa State University Press, 1956.

U. S. Department of Agriculture. "Cost of Food at Home," Family Economics Review (December 1965), 5.

U. S. Department of Agriculture. Food Consumption of Households in the United States, Spring 1965. Household Food Consumption Reports, Nos. 1-10, Washington, D.C., 1968.

Wilson, M. and M. Lamb. "Food Beliefs as Related to Ecological Factors in Women," Journal of Home Economics 60 (February 1968), 115-28.

Young, C. M., B. G. Waldner, and K. Berresford. "What the Homemaker Knows About Nutrition," Journal of the American Dietetic Association, 32 (March 1956), 218-22.

---------------------------------------