Does Food Group Consumption Vary by Differences in Socioeconomic, Demographic, and Lifestyle Factors in Young Adults? The Bogalusa Heart Study

Abstract

Objective

To examine if food group consumption varies by differences in socioeconomic, demographic, and lifestyle factors in young adults from a semirural setting in Louisiana.

Design

Cross-sectional.

Subjects

Young adults (n=1,266, 74% European American, 26% African American; 39% men, 61% women) aged 20 to 38 years, enrolled in the Bogalusa Heart Study.

Measures

Food group consumption was assessed by a food frequency questionnaire. Socioeconomic (eg, income and education), demographic (eg, age, sex, and ethnicity), and lifestyle (eg, marital status and physical activity) information was obtained by a self-administered questionnaire and the subjects were stratified according to these groups.

Statistical analyses

Analysis of covariance (adjusted for covariates) was used to detect differences in the mean servings of food groups consumed per day between the various socioeconomic, demographic, and lifestyle groups.

Results

Compared to income ≤$15,000, those with an income >$45,000 had lower consumption of burgers/sandwiches (P<0.05) and those with income levels from $30,001 to $45,000 had lower consumption of mixed dishes (P<0.05). Intake of cereals/breads (P<0.05), dairy products (P<0.01), fruits/100% fruit juices (P<0.001), and vegetables (P<0.001) was higher in subjects with >12 years of education. European-American men consumed more servings of dairy products (P<0.05) and sweetened beverages (P<0.05) than African-American men. European-American women consumed more servings of dairy products (P<0.05), vegetables (P<0.05), and fats (P<0.05) than African-American women. African Americans (men and women) consumed more servings of fruits/ 100% fruit juices (P<0.0001) than European Americans (men and women), respectively. Married individuals consumed more servings of snacks/desserts (P<0.05), but fewer servings of alcoholic beverages (P<0.0001) than those who were unmarried. Active individuals consumed more servings of fruits/100% fruit juices (P<0.05) and fewer servings of burgers/sandwiches (P<0.05) than inactive individuals.

Conclusions

These findings suggest that food group consumption varies by socioeconomic, demographic, and lifestyle factors in young adults from a semirural setting. Food and nutrition professionals who encounter diverse populations need to consider the influence of income, education, sex, ethnicity, marital status, and physical activity on food consumption patterns when planning diets, nutrition education programs, and interventions for young adults.

With rapid increases in obesity (1), metabolic syndrome (1), cardiovascular disease (2), diabetes mellitus (3), and cancer (4) in the young adult population (ages 19 to 39 years), there is a need for understanding their eating habits and lifestyle patterns. It has been noted that intake of fast foods, snacks (5,6), and sweetened beverages (5) has increased among young adults, whereas intake of fruits, vegetables (7), and dairy products (6) has decreased. Various factors such as socioeconomic status (eg, income and/or education) (8-16), demographics (eg, sex and ethnicity) (17-19), and lifestyle characteristics (eg, marital status [20-23] and physical activity [15,24,25]) have been thought to influence the food choices in this age group. However, the majority of studies conducted in the United States have either examined cost issues with respect to purchasing specific healthful foods (14,26-29) or have examined the influence of socioeconomic status on nutrient intakes (30,31) rather than the consumption of specific food groups. Because foods are usually consumed in many combinations (32), examining the role of single nutrients on socioeconomic status may be confounded by the effect of dietary patterns (33). Also, the high intercorrelation among some nutrients (34) and complicated interactions among nutrients (35) makes it more difficult to assess the influence of socioeconomic status on dietary (nutrient) intakes. Furthermore, reported results of the role of socioeconomic status/cost of foods on dietary intakes have been conflicting, with a majority of the research showing a positive connection (8-12,14-16), and a minority reporting either dubious (13) or no connection (28,29). This warrants more studies.

Ethnic differences in adult nutrient intakes have also been noted (36-38). Yet only a few recent studies have reported such differences in the consumption of specific foods (18-19). Furthermore, past research has depicted inconsistent results; some research shows that European Americans practice more healthful dietary habits than African Americans (17,18) and other research shows mixed results (19). Sex differences in food consumption also prevail; women in general tend to eat more healthfully than men (17). In contrast, it is true that African-American women have the highest rates of overweight/obesity in the United States (39). Thus, it is important to evaluate ethnic and sex differences in food group consumption in light of the current obesity epidemic in the United States.

Marriage has been thought to contribute to healthful living conditions due to sharing of household goods and, hence, reducing the cost of living (21). Yet some research has shown that married individuals have intakes consistent with the dietary guidelines (22) while others have reported that married individuals are more likely to be overweight/obese (40). In contrast, unmarried individuals have been noted to consume poor or less healthful diets (20). Hence, the role of marital status on food group consumption is ambiguous at this juncture.

Although physical activity does not affect the choice of food group consumption directly, it has been noted that unhealthful lifestyle patterns (eg, lack of exercise) and less healthful eating patterns (eg, lower intake of fruits and vegetables and/or higher intake of fats) tend to coexist or cluster among individuals (15,24,25), although more research is warranted in this area, particularly in the young adult population.

Due to the inconsistent findings and focus on nutrient intakes in the past research, there is a need to understand factors that influence food group consumption in young adults residing in semirural communities of the United States. This type of information will help food and nutrition professionals to plan tailored intervention programs to improve the eating habits of young adults relative to food consumption patterns. The purpose of our study was to examine whether food group consumption varies by socioeconomic status (ie, income and education), demographics (ie, sex and ethnicity) and lifestyle factors (ie, physical activity and marital status) in a biracial sample of young adults from the Bogalusa Heart Study.

MATERIALS AND METHODS

Background

The Bogalusa Heart Study began in 1973 as an epidemiologic study investigating the early natural history of cardiovascular risk factors in children; it eventually progressed into observations of young adults (41). Bogalusa is a biracial (European American/white-African American/black) semirural community 70 miles north of New Orleans, LA. The biracial population distribution of Bogalusa has been stable for years and is comparable to many other small southern semirural communities in the United States (41).

Subjects

Our study included a sample of 1,336 young adults aged 20 to 38 years who were examined in the 1995-1996 cross-sectional survey in the Bogalusa Heart Study. The young adults were recruited for this cross-sectional survey on the basis of whether or not they had participated in one of the childhood cross-sectional surveys conducted in the Bogalusa Heart Study. Written informed consents were obtained from all the participants for this study. The study plans, procedures, and consent forms were reviewed and approved by the Louisiana State University and Tulane University medical centers' ethics and research committees.

Procedures

To assess the food consumption patterns of the young adults, a validated (42) and reliable (43) 131-item semiquantitative, self-administered youth and adolescent food frequency questionnaire was used. The participants were asked to indicate how often, on average, they had consumed a given amount of the specified food during the past year. The responses for the foods/beverages (Figure) ranged from “never or less than once a month” to “five or more times a day.” Daily portion sizes were calculated for each of the food items. The selected frequency choice given by the subjects for each food item from the food groups was then converted to a daily intake (eg, “one serving per week” was converted to “0.14 serving/day”). Alcohol consumption was assessed based on the frequency, type, and length of alcohol use. The sum of the daily servings of each of the food/beverage items from a particular food group was used to determine the total number of servings of that food group consumed per day. The total energy intake (in kilocalories) was also computed for all participants.

Figure.

Components of food groups from the youth and adolescent questionnaire: The Bogalusa Heart Study, 1995-1996.

Information on demographics, socioeconomic status, and lifestyle characteristics was obtained using a self-administered questionnaire. The following subcategories were used for classifying the study variables: Age at their most recent birthday; sex (ie, male or female); ethnicity (ie, European American or African American) and ethnicity by sex (ie, European-American man or European-American woman or African-American man or African-American woman). Socioeconomic status of the young adults was evaluated by eliciting information on their income levels (ie, ≤$15,000, $15,001 to 30,000, $30,001 to 45,000, >$45,000) and their education levels (ie, ≤12 years or >12 years). For information on their marital status, young adults were asked if they were currently married/cohabiting, or single (which included divorced and widowed). Physical activity level of the participants was also assessed using a self-report method. The participants were asked to give a subjective rating of their physical activity outside of work following a five-item Likert scale adapted from the Lipid Research Questionnaire (44), where 1=physically inactive (sedentary) and 5=very active. The participants were considered inactive if they classified themselves as ≤3 on the Likert scale and active if they classified themselves as ≥4 on the Likert scale.

Trained examiners measured heights and weights of participants following standardized protocols and using calibrated instruments (41). Height and weight were measured twice to the nearest 0.1 cm and 0.1 kg, respectively, using a standard height board and beam balance metric scale. The two readings were averaged for both height and weight. Body mass index (BMI) was calculated as kg/m².

Statistical Analyses

Statistical analyses were performed using the Statistical Analysis System (version 8.2, 1999, SAS Inc, Cary, NC). Descriptive statistics (ie, means, standard deviations/ standard errors, and percentages) were computed wherever necessary. t-tests and analysis of variance (with Tukey-Kramer's post hoc tests) were used to detect ethnicity, sex, and ethnicity by sex differences among the demographic variables, respectively. χ² analysis was used for testing the associations among the categorical demographic variables. Analysis of covariance was used to detect differences in the mean servings of food groups consumed per day among the various socioeconomic, demographic, and lifestyle groups. The dependent variable was the mean servings of the food groups consumed per day and the various independent variables were income, education, sex, ethnicity, ethnicity by sex, marital status, and physical activity. Because Kendall Tau-β coefficients showed significant correlations between income levels, education levels, marital status, and physical activity, these variables were included in the same model to control for each other, along with the other covariates (ie, energy intake, BMI, age, sex, ethnicity, and ethnicity by sex). Tukey-Kramer's post hoc significance tests were applied for multiple comparisons between any two groups. The statistical significance was set at P<0.05.

RESULTS

From a total of 1,336 subjects, 70 subjects were excluded from the analyses as outliers, which included women whose daily energy intakes >4,000 kcal/day (n=28) or those who were pregnant (n=24), and men with daily energy intakes >4,200 kcal/day (n=18). None of the subjects consumed extremely low energy intake (<600 kcal/day) or were lactating in this sample.

Demographic Characteristics of the Sample

As shown in Table 1, the cross-sectional sample of 1,266 young adults consisted of 74.4% European Americans, 25.6% African Americans, 38.7% men, and 61.3% women. During the analysis phase, the sample size varied from the original sample size due to missing data for income, education, and physical activity for some participants. The mean age of subjects was 29.7 years. There were significant ethnicity by sex differences in BMI, energy intake, income levels, education levels, and physical activity levels among the young adults. In short, African-American women had higher BMIs, African-American men had higher energy intakes, and European-American women had lower BMIs and lower energy intakes than all the other ethnicity by sex groups. Furthermore, a higher percent of African Americans reported having lower income levels and a higher percent of European Americans reported having >12 years of education. African-American men were the most physically active among all the ethnicity by sex groups and a higher percent of European Americans reported being married than did African Americans.

Table 1.

Demographic characteristics^a of young adults by ethnicity and sex: The Bogalusa Heart Study, 1995-1996

Characteristics of young adults	Ethnicity		Sex		Ethnicity by Sex
	EAb	AAc	Males	Females	EA males	EA females	AA males	AA females	Total
Original sample size
n	942	324	490	776	368	574	122	202
%	74.4	25.6	38.7	61.3	29.1	45.3	9.6	16.0	1,266
Body mass index (mean±SDd)	26.7±6.2	29.3 ±8.1****	27.8±5.8	27.1 ±7.4	27.8±5.2x	26±6.7y	27.8±7.3x	30±8.4z	27.4±6.8
Age (mean±SD)	29.9±5	29.2±5.3*	30.2±5.1	29.4±5.1**	30.3±5x	29.6 ±5xy	29.9±5.3xy	28.7±5.3y	29.7±5.1
Energy intake (kcal) (mean±SD)	2,055±636.8	2,290.7±759.4****	2,262±722.3	2,023.5±631.4****	2,208.8±687.6x	1,957.5±581.7y	2,422.5±799.9z	2,211.0±724.3x	2,115.8±677.8
Income levelse
≤$15,000
n	203	176	135	244	69	134	66	110	379
%	23	63	30	33	20	24	65	61	32
$15,001-$30,000
n	255	63	128	190	104	151	24	39	318
%	28	23	28	26	29	27	24	22	27
$30,001-$45,000
n	200	27	83	144	79	121	4	23	227
%	22	10	18	20	22	22	4	13	19
>$45,000
n	246	14	108	152	101	145	7	7	260
%	27	5	24	21	29	26	7	4	22
Educatione >12 y
n	629	151	272	508	226	403	46	105	780
%	72	54	61	71	66	75	43	60	67
Educatione ≤12 y
n	250	130	176	204	116	134	60	70	380
%	28	46	39	29	34	25	57	40	33
Physical activityef Mean±SD	3.2±1	3.4±1.2***	3.4±1.1	3.1±1.1****	3.4±1.0x	3.1±1.1y	3.7±1.3z	3.2±1.2xy	3.2±1.1
Married young adults
n	568	75	233	410	209	359	24	51	643
%	61	23	48	53	57	63	20	25	51

^aIncome levels were associated with ethnicity (P<0.0001) and ethnicity by sex (P<0.0001); education levels were associated with ethnicity (P<0.0001), sex (P<0.0001), and ethnicity by sex (P<0.0001). Being married was associated with ethnicity (P<0.0001). Values with different superscripts (x, y, z) have significant mean differences at P<0.0001. Pairwise comparisons tested by Tukey-Kramer's multiple comparison test are significant at P<0.05.

^bEA=European Americans.

^cAA=African Americans.

^dSD=standard deviation.

^eSample size varies from original due to missing data for income levels (n=82), education levels (n=106), and physical activity levels (n=75).

^fLikert scale (1-5): 1 to 3=physically inactive (sedentary); 4 to 5=physically very active.

^*P<0.05.

^**P<0.01.

^***P<0.001.

^****P<0.0001.

Socioeconomic Status and Food Group Consumption

As shown in Table 2, compared with those who reported lower income levels, young adults with higher income levels consumed significantly fewer servings of burgers and sandwiches and mixed dishes. Similarly, those young adults with a reported education level >12 years consumed significantly more servings of breads and cereals, dairy products (eg, cheese and yogurt), fruits, 100% fruit juices, and vegetables (with and without including french fries) than those with an education level of ≤12 years (Table 2).

Table 2.

Food group consumption^a in young adults (aged 20 to 38 years) by socioeconomic status: The Bogalusa Heart Study, 1995-1996

Food Groups	Income Levels (n=1,184)b				Education Levels (n=1,160)c
	≤$15,000 (n=379)	$15,001-30,000 (n=318)	$30,001-45,000 (n=227)	>$45,000 (n=260)	≤12 y (n=380)	>12 y (n=780)
Breads and cereals	2.18	2.26	2.23	2.28	2.16	2.32*
Milk	0.61	0.64	0.60	0.70	0.60	0.68
Cheese	0.36	0.36	0.36	0.34	0.32	0.39*
Yogurt	0.008	0.038	0.035	0.04	0.02	0.04*
Dairy products	0.98	1.04	1.00	1.08	0.94	1.11**
Fruits and 100% fruit juices	1.10	1.26	1.30	1.35	1.14	1.37***
Vegetables including french fries	2.12	2.30	2.36	2.34	2.15	2.41***
Vegetables excluding french fries	1.91	2.08	2.13	2.12	1.91	2.2***
Fats	1.80	1.54	1.52	1.44	1.57	1.58
Meats	1.12	1.16	1.16	1.15	1.15	1.14
Burgers and sandwiches	0.78x	0.74xy	0.76x	0.68y	0.76	0.73
Mixed dishes	0.56x	0.50y	0.49y	0.51xy	0.5	0.53
Snacks and desserts	2.21	2.19	2.32	2.51	2.36	2.25
Sweetened beverages	1.60	1.54	1.48	1.41	1.6	1.45
Alcoholic beverages	0.19	0.17	0.2	0.21	0.19	0.20

^aAll values are least squared means of food group servings consumed per day. Standard errors (not shown) are available from the authors upon request.

^bSample size varies from original due to missing data (n=82). Analysis of covariance model for income levels adjusted for energy intake, body mass index, age, education, marital status, physical activity level, ethnicity, sex, and ethnicity by sex. Different superscripts (x, y) indicate significant mean differences at P<0.01 for both burgers and sandwiches and mixed dishes, respectively. Pairwise comparisons tested by Tukey-Kramer's multiple comparison test are significant at P<0.05.

^cSample size varies from original due to missing data (n=106). Analysis of covariance model for education levels adjusted for energy intake, body mass index, age, income, marital status, physical activity level, ethnicity, sex, and ethnicity by sex.

^*P<0.05.

^**P<0.01.

^***P<0.001.

Demographic Factors and Food Group Consumption

In general, men consumed significantly more servings of burgers and sandwiches and alcoholic beverages than women (Table 3). In contrast, women consumed significantly more servings of yogurt, fruits and 100% fruit juices, vegetables (with and without including french fries), mixed dishes, and fats than did men. Ethnic differences in food group consumption revealed that European-American young adults consumed significantly more servings of dairy products (eg, milk and cheese), vegetables (with and without including french fries), fats, mixed dishes, and sweetened beverages than their African-American counterparts. In contrast, African-American young adults consumed significantly more servings of fruits, 100% fruit juices, snacks, desserts, and alcoholic beverages than their European-American counterparts.

Table 3.

Food group consumption^a in young adults (aged 20 to 38 years) by ethnicity and sex: The Bogalusa Heart Study, 1995-1996

Food group	Ethnicity (n=1,266)		Sex (n=1,266)		Ethnicity by Sex (n=1,266)
	EAb (n=942)	AAc (n=324)	Males (n=490)	Females (n=776)	EA males (n=368)	EA females (n=574)	AA males (n=122)	AA females (n=202)
Breads and cereals	2.26	2.3	2.24	2.28	2.2	2.29	2.35	2.26
Milk	0.8	0.58****	0.79	0.73	0.83z	0.78y,z	0.6xy	0.57x
Cheese	0.43	0.37*	0.39	0.43	0.4xy	0.45y	0.39xy	0.36x
Yogurt	0.04	0.04	0.03	0.05*	0.03	0.05	0.03	0.04
Dairy products	1.28	0.98****	1.21	1.2	1.27y	1.29y	1.01x	0.97x
Fruits and 100% fruit juices	1.14	1.53****	1.17	1.28*	1.08y	1.18y	1.45x	1.58x
Vegetables including french fries	2.5	2.13****	2.21	2.52****	2.29z	2.64y	1.97x	2.21xz
Vegetables excluding french fries	2.27	1.91****	1.92	2.26***	2.06	2.47	1.78	2.04
Fats	1.83	1.36****	1.49	1.85****	1.57z	2y	1.22x	1.43xz
Meats	1.11	1.13	1.1	1.12	1.09	1.12	1.41	1.13
Burgers and sandwiches	0.73	0.76	0.82	0.69****	0.81y	0.68x	0.86y	0.7x
Mixed dishes	0.56	0.49****	0.53	0.55*	0.53x	0.58y	0.5x	0.49x
Snacks and desserts	2.14	2.42**	2.16	2.25	2.09y	2.18xy	2.38xy	2.44x
Sweetened beverages	1.65	1.48**	1.61	1.6	1.71y	1.61y	1.32x	1.58xy
Alcoholic beverages	0.17	0.24****	0.25	0.16***	0.20	0.14	0.29	0.18

^aAll values are least squared means of food group servings consumed per day. Standard errors (not shown) are available from the authors upon request. Analysis of covariance models for sex, ethnicity, and ethnicity by sex adjusted for energy intake, body mass index, and age, respectively. Different superscripts (x, y, z) indicate significant mean differences for milk (P<0.001), cheese (P<0.05), dairy products (P<0.0001), fruits and 100% fruit juices (P<0.0001), vegetables, including french fries (P<0.0001), fats (P<0.0001), burgers/sandwiches (P<0.0001), mixed dishes (P<0.0001), snacks and desserts (P<0.01), and sweetened beverages (P<0.01). Pairwise comparisons tested by Tukey-Kramer's multiple comparison test are significant at P<0.05.

^bEA=European Americans.

^cAA=African Americans.

^*P<0.05.

^**P<0.01.

^***P<0.001.

^****P<0.0001.

Post hoc comparisons (Table 3) for the ethnicity by sex groups revealed that European-American men consumed significantly more servings of: milk than did African-American men and African-American women, vegetables (including french fries) than did African-American men, fats than did African-American men, burgers and sandwiches than did European-American women and African-American women, and sweetened beverages than did African-American men. In contrast, African-American men consumed significantly more servings of: fruits and 100% fruit juices than did European-American men and European-American women; and burgers and sandwiches than did European-American women and African-American women.

European-American women consumed significantly more servings of: dairy products (eg, milk and cheese) than did African-American women; vegetables (including french fries) than did European-American men, African-American men, and African-American women; fats than did European-American men, African-American men, and African-American women; mixed dishes than did European-American men, African-American men and African-American women; and sweetened beverages than did African-American men. On the other hand, African-American women consumed significantly more servings of: fruits and 100% fruit juices than did European-American men and European-American women, snacks and desserts than did European-American men.

Lifestyle Factors and Food Group Consumption

The number of servings of snacks and desserts was significantly higher in married young adults compared with unmarried young adults (Table 4). In contrast, unmarried young adults consumed significantly more servings of alcoholic beverages than married young adults. The number of servings of burgers and sandwiches was significantly higher in the inactive young adults than in those who were active. The intake of fruits and 100% fruit juices was significantly higher in active individuals than in inactive individuals (Table 4).

Table 4.

Food group consumption^a in young adults (aged 20 to 38 years) by lifestyle factors: The Bogalusa Heart Study, 1995-1996

Food group	Marital Statusb (n=1,266)		Physical Activity Levelsc (n=1,191)d
	Married (n=643)	Not married (n=623)	1 (n=74)	2 (n=181)	3 (n=544)	4 (n=180)	5 (n=212)
Breads and cereals	2.26	2.22	2.11	2.37	2.17	2.36	2.19
Milk	0.65	0.62	0.54	0.64	0.61	0.65	0.76
Cheese	0.35	0.36	0.31	0.36	0.39	0.37	0.34
Yogurt	0.02	0.04	0.01	0.03	0.03	0.03	0.05
Dairy products	1.03	1.02	0.86	1.02	1.04	1.05	1.15
Fruits and 100% fruit juices	1.22	1.29	1.14xy	1.06x	1.27xy	1.37y	1.42y
Vegetables including french fries	2.21	2.35	2.17	2.18	2.24	2.38	2.43
Vegetables excluding french fries	2.00	2.12	1.93	1.95	2.03	2.18	2.21
Fats	1.62	1.53	1.73	1.62	1.54	1.53	1.46
Meats	1.14	1.15	1.24	1.11	1.11	1.11	1.16
Burgers and sandwiches	0.75	0.74	0.72xy	0.78x	0.75xy	0.69y	0.77xy
Mixed dishes	0.52	0.51	0.56	0.51	0.52	0.50	0.49
Snacks and desserts	2.40	2.21*	2.18	2.40	2.37	2.26	2.32
Sweetened beverages	0.60	0.68	1.60	1.47	1.54	1.57	1.37
Alcoholic beverages	0.16	0.23****	0.17	0.18	0.21	0.2	0.22

^aAll values are least squared means of food group servings consumed per day. Standard errors (not shown) are available from the authors upon request.

^bAnalysis of covariance model for marital status adjusted for energy intake, body mass index, age, income, education, ethnicity, sex, ethnicity by sex, and physical activity.

^cLikert scale (1-5): 1 to 3=physically inactive (sedentary), 4 to 5=physically very active. Analysis of covariance model for physical activity levels adjusted for energy intake, body mass index, age, income, marital status, education, ethnicity, sex, and ethnicity by sex. Different superscripts (x, y) indicate significant mean differences for fruits and 100% fruit juices (P<0.01) and burgers and sandwiches (P<0.05). Pairwise comparisons tested by Tukey-Kramer's multiple comparison test are significant at P<0.05.

^dSample size varies from original due to missing data for physical activity (n=75).

^*P<0.05.

^****P<0.0001.

DISCUSSION

It has been well established that increased consumption of healthful foods (eg, whole grains/cereals [45], fruits, vegetables [45], low-fat dairy products [46], and other low-fat foods [47]) is associated with lower mortality rates from obesity and related metabolic diseases (45-47). These benefits have been mainly attributed to their vitamin/mineral (46,48), phytochemical (48), and fiber content (49). Conversely, higher intakes of high-fat foods or foods that provide discretionary calories (50) (eg, fast foods [1], snacks [51], desserts [4], sugar-sweetened beverages [3], and alcohol [52]) have been proposed to be associated with obesity and related metabolic diseases.

Findings from our research reveal that food group consumption varies by socioeconomic, demographic, and lifestyle factors in young adults. It was noted that young adults with higher income levels consumed fewer servings of burgers and sandwiches and mixed dishes and those with higher education levels consumed more servings of healthful foods (eg, breads and cereals, dairy products, fruits, 100% fruit juices, and vegetables). Similar findings from previous research have been reported; individuals with a lower socioeconomic status had a higher consumption level of potatoes, meat products, visible fats, and soft drinks (11); another study reported higher consumption of chips, soda, white bread, processed meats, full-fat dairy products, and sugar in women with a lower socioeconomic status (15). In contrast, individuals belonging to higher socioeconomic status groups were observed to have more consumption of whole grains (15), fruits/vegetables (11,15), and dairy products (11-13,15). It has been hypothesized that the relationship between higher socioeconomic status and more healthful food consumption could be due to the knowledge and health awareness that individuals may acquire with higher socioeconomic status, or increased pressures of social acceptability that occur with increasing socioeconomic status and these factors may influence their food consumption habits (53).

Our finding of socioeconomic status influencing food group consumption is interesting in light of the age group chosen for this study (ie, young adults). It has been reported that individuals belonging to younger age groups prefer to spend less time in food preparation (54). Furthermore, younger age groups may be more inclined to prefer ready-to-eat, processed, frozen, or canned fruits and vegetables for convenience. Yet, the high costs of some of these nutritious foods owing to their high processing costs and/or their highly perishable nature may be a strong impediment to eating healthfully (26,27) and, thus, is proposed to be a major factor mediating the relationship between socioeconomic status and healthful food consumption. One study reported that adolescents tend to eat more fast foods because they think that fast foods are inexpensive and easily available (54) compared to healthful foods. Despite the findings of our study and from past research (11,12,15,26,27), a recent economic analysis report from the US Department of Agriculture found that three servings of fruits and four servings of vegetables can be easily bought at a low cost of 64 cents daily (29). However, their findings did not consider the perishable nature of fruits and vegetables (29), which typically tends to affect the purchasing ability of these foods. More research is clearly warranted in this area and nutrition education programs should target cost issues in conjunction with attempts to improve the dietary habits of young adults.

The results for sex and ethnic differences in food group consumption in our study were mixed, yet interesting. Men consumed more servings of burgers and sandwiches. Conversely, women consumed more servings of healthful foods (eg, yogurt, fruits, 100% fruit juices, and vegetables). Similar findings have been reported earlier: One study found that women consumed more servings of fruits and vegetables and fewer servings of meat (17). On the other hand, we also noted that women consumed more servings of fats and mixed dishes, and these results are contradictory to previous studies (17) and it may be possible that our study findings may be specific only to the Bogalusa population.

Analogous results to our study have been depicted for ethnic differences in food consumption. For example, Champagne and colleagues (19) found that discretionary fat consumption was higher and fruits, grains, and cereal consumption were lower in European Americans when compared with African Americans. In contrast, another study reported that European Americans were more likely to use low-fat alternatives and replacements for meat in their diets than were African Americans (18). This discrepancy in results could be due to socioeconomic and demographic differences among the populations studied.

Men and African-American young adults were the groups that consumed more servings of alcoholic beverages in our study. Similar sex differences have been noted elsewhere (17,55). However, contrasting results have been noted for ethnic differences in alcohol intakes: one study found similar results to ours (56), another study found that European Americans consumed more alcohol than did African Americans (55). Reasons for this variation in results may be due to differences in sample size and age of the subjects, differences in definitions of alcohol consumption, underreporting of alcohol consumption by the subjects (57), or differences in the statistical methods used.

The basis for ethnic disparities in food consumption could be attributed to the fact that a larger number of African Americans were in the lower income group than European Americans (as observed in our study and also in previously published studies [58]) and hence may have consumed poorer diets (19) due to their inability to afford more healthful foods. Some other environmental factors like geographic reasons may also be responsible for these ethnic differences in food consumption. One previous study showed that predominantly low-income African-American neighborhoods in Louisiana had 2.4 fast food restaurants per square mile compared to 1.5 fast food restaurants in European American neighborhoods (59).

In our study married young adults consumed more servings of snacks and desserts and fewer servings of alcoholic beverages than those who were unmarried. It has been reported previously that married individuals tend to eat a greater number of meals and have higher energy intakes (23). These results from our study and previous studies (23) may be a reason for the increased weight gain among married individuals as reported in the literature (40).

Some previous studies have also reported that married individuals tend to eat more healthful diets than unmarried individuals (22). However, such differences were not found in our study. One can speculate that with both spouses working there may be fewer family-oriented meal consumption patterns and hence being married did not affect the food consumption patterns among these young adults. With respect to alcohol consumption and marital status, previous studies have reported contradictory results to ours. One study showed that married individuals consumed more alcohol than did unmarried individuals (55). Our findings may be specific to the Bogalusa population. Nevertheless, further investigation is essential in this area.

In our study less-active individuals consumed more servings of burgers and sandwiches and fewer servings of fruits and 100% fruit juices than did physically active young adults. One can hypothesize that such findings could be due to behavioral factors (60) (ie, individuals who are physically active could be more health conscious and may choose to consume more healthful foods). However, studies have not been able to establish a direct biologic link between increased physical activity and increased fruit consumption. Further, our study did not find statistically significant differences in consumption of vegetables and dairy products with respect to physical activity. Although not many studies have examined this issue, Jago and colleagues (25) reported that subjects who were more physically active consumed more servings of fruits, 100% fruit juices, and dairy products per day and a lesser percent of energy from fats than those who were less active. Another study (24) reported that a so-called cosmopolitan diet, mainly distinguished by higher consumption of vegetables, rice, pasta, vegetable oil, chicken, fish, and wine and lower intake of potatoes, was strongly associated with leisure-time physical activity. From these findings one can hypothesize that less-healthful food consumption patterns tend to occur with physical inactivity in young adults, although more studies are warranted in this area.

Our research does have some limitations. First, the cross-sectional design provided only a snapshot of the influences of socioeconomic, demographic, and lifestyle factors on the food group consumption patterns of young adults; therefore, causal links cannot be established (61). Second, despite the large biracial sample included in this study, the breakdown of the sample size inadvertently underrepresented the African-American young adults. Third, the researchers used youth/adolescent questionnaire instead of the adult version of the food frequency questionnaire (62) for dietary assessment. A major difference between the two questionnaires is that the youth/adolescent questionnaire contains more snack foods. Furthermore, a review of 24-hour dietary recalls in a previous Bogalusa cross-sectional survey revealed that the young adults consumed a lot of snack foods (63). Also, the total energy intake and percentage of energy from fat as measured using the youth/adolescent questionnaire from Bogalusa (64) was similar to other studies using the 24-hour dietary recall method from Bogalusa with the same age group (63) and to the 1995 Continuing Survey of Food Intakes by Individuals measured by a 2-day dietary recall (65). Collectively, these points provide a strong rationale for the use of youth/adolescent questionnaire for the young adult age group in this study. Fourth, this study relied on self-reported information from the participants. Dietary information obtained from a self-reported food frequency questionnaire may have substantial errors (eg, difficulty in the estimation of portion sizes among persons [66], recall bias for food consumption [66], and under- or over-reporting for less healthful or more healthful foods, respectively [57,66,67]). The influence of self-report is minimized for the measure of physical activity because past studies have shown that the five-item physical activity scale we used has high accuracy, with a test–retest reliability of 0.85, and has been reported to be significantly associated with 4-week physical activity (44). Nevertheless more objective measures of physical activity (eg, heart rate monitors or physical activity monitors [68]) could have added more credibility to the methodology. Fifth, the classification of young adults into only two marital status categories (ie, married/cohabiting and single individuals, which included divorced and widowed) for the ease of statistical analyses, could have overlooked certain issues (eg, married and cohabiting young adults could have different eating patterns, single individuals such as unmarried divorced and widowed persons could also differ in their eating habits, and the presence of children in the household could also mediate the relationship between food group consumption and marital status). Despite all these shortcomings, this study makes a valuable contribution to the literature on factors that influence food group consumption in young adults from a southern semirural community in the United States.

CONCLUSIONS

The results of our study suggest that food group consumption varies by socioeconomic, demographic, and lifestyle factors in young adults from a semirural community in Bogalusa, LA. Public health research nutritionists and other food and nutrition professionals who encounter diverse populations need to consider the influence of income, education, sex, ethnicity, marital status, and physical activity on food consumption patterns when planning diets, nutrition education programs, and interventions for young adults. Furthermore, data suggests that intervention programs may need to be tailored based on socioeconomic, demographic, and lifestyle differences in food consumption patterns. However, more studies are needed to confirm these findings in other young adult populations.