- Research article
- Open Access
- Open Peer Review
Association of household income and education with eating behaviors in Japanese adults: a cross-sectional study
BMC Public Health volume 16, Article number: 61 (2016)
Socioeconomic inequalities as social determinants of health are important issues in public health and health promotion. However, the association between socioeconomic status and eating behaviors has been investigated poorly in Japanese adults. To fill this gap, the present study examines the association of eating behaviors with household income and education.
The sample comprised 3,137 Japanese adults (1,580 men and 1,557 women) aged 30 to 59 years who responded to an Internet-based cross-sectional survey in 2014. Data on the following eating behaviors were collected via self-report: “taking care of one’s diet for health,” “eating vegetables,” “frequency of eating breakfast,” “frequency of family breakfasts,” “frequency of family dinners,” “using the information on nutrition labels,” and “conversations with family or friends during meals.” Self-reported data on socioeconomic status (household income and education) and demographic variables (gender, age, district of residence, marital status, residence status, and employment status) were also collected. The associations between eating behaviors and household income or education were tested using binomial logistic regression analysis with eating behaviors as dependent variables and household income and education as independent variables. A trend P -value was calculated for three categories of household income (less than 3,000,000 JPY, 3,000,000–7,000,000 JPY, and over 7,000,000 JPY) and education (junior high/high school, 2-year college, and 4-year college/graduate school).
Higher household income and education were significantly associated with higher rates of eating vegetables, using the information on nutrition labels, and conversation with family or friends during meals in Japanese men and women. Higher household incomes were significantly associated with lower rates of frequency of family breakfasts in Japanese men and lower rates of frequency of family dinners in Japanese men and women.
Higher socioeconomic status as indicated by household income or education was associated with eating more vegetables and conversation with family or friends during meals in Japanese men and women. Socioeconomic status should be considered in health promotion and diet improvement.
Health disparities are important issues in public health and health promotion. Socioeconomic disparities in health have been widely reported in many Western countries. In several previous studies, individual socioeconomic status (SES), as measured by household income, education, or occupation, has been shown to be closely related to lifestyle, mortality, and morbidity [1, 2]. In Japan, reduction of health disparities caused by differences in SES is considered important for public health and health promotion [3, 4]. Until recently, few major socioeconomic disparities have been reported among the Japanese population . Several recent reports, however, have shown that a lower SES is generally associated with higher likelihood of behaviors with a health risk [6, 7] as well as higher mortality and morbidity [8–11]. A review of the research on health differences will therefore support measures to reduce health disparities in Japan.
The World Health Organization (WHO)  identified the main intermediary determinants of health, usually behavioral or social factors, as nutritional and dietary behaviors, physical activity, tobacco consumption, and alcohol consumption. These factors differ by SES; dietary habits, for example, are an important factor in disease prevention and health promotion among those with low SES [13–15]. Fewer of those with a household income of less than 2,000,000 JPY have an adequate vegetable intake than those whose household income is more than 6,000,000 JPY . Murakami  reported that lower education levels and working outside the home were associated with an unfavorable dietary intake pattern in a group of pregnant Japanese women. Fukuda  reported that lower household expenditure was associated with unhealthy and unbalanced nutrient intake in Japanese adults. These differences in nutritional intake were also predicted by sex, age, and marital status. Health status and nutrition are affected by food intake, which is in turn affected by eating behaviors. The relationship between SES and dietary behavior has a number of known mediators and moderators , but the relationships between eating behaviors and SES in Japan are largely unexplored. As a mechanism to transform dietary behavior, we applied the Healthy Japan 21 policy framework of goal-setting for eating habits [12, 19].
Dietary behavior is affected by knowledge, skills, attitudes, food preparation factors and environment . Nutrition education aims to transform eating behavior. It is necessary to measure socioeconomic factors, such as household income or education, to promote desirable action and transformation . By examining the association between healthy eating behaviors and SES, we can set SES-specific action targets to support changes in eating patterns. There are, however, very few studies on the association between household income or education and eating behavior , although it has been reported that the frequency of eating breakfast differs by sex and household income . To obtain basic data as a foundation for research on health disparities, it is first necessary to establish whether healthy eating behaviors differ by sex, as in previous studies [17, 18], and to examine the association between SES (household income and education) and healthy eating behavior. If we can establish the relationship between dietary behavior and SES, this will inform the development of future SES-based interventions.
We hypothesized that people with higher SES would have healthier eating habits, which would in turn influence their general health, because higher SES permits adequate intake of healthy food and therefore good health. We used seven items chosen from the Japan National Health and Nutrition Survey (NHNS)  and previous studies [20–30]. Other projects include a child poverty program  and the Japan Gerontological Evaluation Study (JAGES) . As adults are responsible for raising the next generation, we decided that targeting adults would have a longer term effect on reducing health disparities. We therefore examined differences in eating behaviors among adults with different SES. Associations between SES and eating behaviors in Japan may differ from those found in other countries because of variations in socioeconomic conditions. We therefore examined the association between SES and healthy eating behaviors in a sample of Japanese adults, to provide support for interventions to improve dietary habits in Japan.
An Internet-based cross-sectional survey was conducted in February 2014 by a Japanese online research service company that holds data on approximately 160,300 adult registrants, including their sociodemographic attributes. This allows the company to target particular demographic groups as necessary. We aimed to have a sample of approximately 3,000 adults aged 30 to 59 years, with 500 men and 500 women from three age groups (30–39, 40–49, and 50–59 years). We targeted adults aged 30 to 59 because we felt that both promotion of healthy eating and reduction in health disparities were particularly important in this group. The set sample size and attributes were stratified by distribution of the Japanese average age, the 2013 Population Census of Japan for sex , and the 2012 Comprehensive Survey of Living Conditions in Japan for household income . In total, 8,284 adults were randomly selected from the database and received an e-mail inviting them to participate in our survey. The invitation e-mail contained a URL directing potential respondents to a protected area of the website where the questionnaire was located. They could then log on using their ID and password. The research service company offered rewards points valued at 100 JPY (one USD was equivalent to approximately 102 JPY in February, 2014). Of those invited, 3,269 answered the survey questions online (a response rate of 39.5 %). Respondents who completed the questionnaire and clicked on the Send button at the end of the online informed consent form were considered to have agreed to participate in the survey. The study received prior approval from the Ethics Review Committee on Research with Human Subjects of Waseda University, Japan.
Socioeconomic status and sociodemographic variables
Household income and educational level were used to measure SES. Educational level was grouped into three categories: junior high/high school, 2-year college, and 4-year college/graduate school. We were unable to accurately ascertain individual-level equivalent incomes because the choices were for levels of household income: less than 3,000,000 JPY, 3,000,000–5,000,000 JPY, 5,000,000–7,000,000 JPY, 7,000,000–10,000,000 JPY, 10,000,000–15,000,000 JPY, and over 15,000,000 JPY. We therefore analyzed income at category level. Household income was classified into three equally distributed categories: less than 3,000,000 JPY, 3,000,000–7,000,000 JPY, and over 7,000,000 JPY.
Seven items were used to measure eating behavior: taking care of one’s diet for health, eating vegetables [20, 21], frequency of eating breakfast [22, 23], frequency of family breakfasts and dinners [24–26], using the information on nutrition labels [27, 28], and conversations with family or friends during meals [29, 30]. These dietary behaviors were in line with the aims of programs like Healthy Japan 21. Questions about eating behaviors were preceded by the phrase, “The following questions are about your normal meals”. Respondents were asked, “Do you normally take an interest in nutrition and healthy meals?” There were six response choices: (1) Very often; (2) Often; (3) Sometimes; (4) Rarely; (5) Almost never; and (6) Never. Participants who answered (1) to (3) for this question were defined as those taking care of one’s diet for health. Respondents were asked, “Do you eat adequate amounts of vegetables (5 small dishes/day, or about 350 g)?” Responses included the following four choices: (1) Very often; (2) Often; (3) Not much; and (4) Never. People who answered (1) or (2) to this question were defined as those who eat sufficient amounts of vegetables.
The next questions were preceded by the phrase, “The following questions are about your normal lifestyle”. Respondents were asked, “How often do you usually eat breakfast?”, with five possible responses: (1) Every day; (2) 4 or 5 days/week; (3) 2 or 3 days/week; (4) 1 day/week; and (5) Never. Participants who answered (1) to this question were defined as those who eat breakfast regularly. Respondents were asked, “How often each week do you usually eat breakfast with all the members of your family?” and “How often each week do you usually eat dinner with all the members of your family?” The following five response choices were provided: (1) Every day; (2) 4 or 5 days/week; (3) 2 or 3 days/week; (4) 1 day/week; and (5) Never. People who answered (1) were defined as those who regularly eat meals with their families each question.
The final questions were preceded by, “The following questions are about your normal eating habits”. Respondents were asked, “Do you use the information on nutrition labels or calorie information on store displays and menus?” Possible responses were: (1) Very often; (2) Often; (3) Not much; and (4) Never. Participants who answered (1) or (2) were defined as those who consult nutrition information. Respondents were asked, “Do you talk to your family and friends during meals about the meal or nutrition?” with four response choices: (1) Very often; (2) Often; (3) Not much; and (4) Never. People who answered (1) or (2) to this question were defined as having positive conversations during meals.
Health risk behaviors
The study asked about two health-risk behaviors: current smoking and alcohol consumption. Smoking habits were surveyed with the question, “How many cigarettes or cigars per day do you smoke?” Responses were: (1) I have never smoked; (2) I stopped smoking more than 1 year ago; (3) I stopped smoking less than 1 year ago; (4) I smoke 1–20 cigarettes or cigars per day (5) I smoke 21–40 cigarettes or cigars per day; and (6) I smoke more than 41 cigarettes or cigars per day. We categorized participants who responded (4) to (6) as current smokers. Alcohol consumption was surveyed by asking, “How many days per week do you consume alcohol?” Response choices were as follows: (1) Every day; (2) 5 or 6 days/week; (3) 3 or 4 days/week; (4) 1 or 2 days/week; (5) 1 to 3 days/month; (6) I stopped consuming alcohol more than 1 year ago; and (7) I hardly drink alcohol at all. We categorized responses (1) to (5) as current alcohol consumption.
Demographic variables included sex, age, marital status, residence status, and employment status. Age was classified as 30–39, 40–49, and 50–59 years. Marital status was categorized as currently married or currently unmarried. Residence status was categorized as living with others or living alone. Employment status was categorized as employed or not employed.
Data were analyzed for the 3,137 adults who provided complete information for the study variables. Respondents who did not provide education status (Other/Unknown, n = 52) or employment status (Other/Unknown, n = 80) were not included in the analysis because these were important variables in this study. We considered responses of Other/Unknown to be missing values rather than lost data. Interpretation of the results would have been difficult if Other/Unknown responses were combined with the other choices, not least because there were very few such responses. We therefore excluded these data from the analysis.
Statistical analysis was performed overall and separately by sex. The chi-squared test was used to compare various characteristics and eating behaviors between men and women. Associations between SES and healthy eating behaviors were examined using forced-entry adjusted logistic regression analysis. Unadjusted odds ratios (OR), adjusted odds ratios (AOR), and 95 % confidence intervals (CI) were calculated for each variable. Associations between eating behaviors and household income or education were determined using binomial logistic regression analysis, with eating behaviors as the dependent variable and household income and education as independent variables. Household income as an independent variable was adjusted for age group; marital, residence and employment status; and education. Education as an independent variable was adjusted for age group; marital, residence and employment status; and household income. Previous studies have found that differences in eating behaviors were predicted by sex, age, marital status, and residence status [6, 17, 18], and have also adjusted for sex, age, and residence status . We also adjusted for employment status for two reasons, first, because nutrient intake and being employed have previously been shown to be related in women , and second because being employed influences household income and its relationships to dietary behavior. We did not adjust for alcohol consumption and smoking because the associations did not change when adjusted. A trend p-value was calculated for the three categories of household income and education, and p values of < 0.05 were considered significant. IBM SPSS Statistical Package for Windows Version 21.0 (IBM Japan Inc., Tokyo, Japan) was used for all statistical analyses.
Table 1 shows demographic characteristics of the 3,137 respondents. Overall mean age was 44.1 (SD = 8.1) years and was approximately the same for men and women. Substantial proportions were current smokers and consumers of alcohol (28.0 % and 72.3 % of men, 13.5 % and 50.2 % of women).
The prevalence of each eating behavior is shown in Table 2. Nearly 75 % of participants reported a positive intention to take care of their diet for the sake of their health. The frequency of eating breakfast was higher frequency. Most participants reported a lower frequency of eating vegetables, family breakfasts, family dinners, using nutrition information, and positive conversations with family or friends during meals. The eating behaviors of men and women differed significantly.
Table 3 shows the results of logistic regression analysis of the association between household income and eating behaviors. In the unadjusted analysis, taking care of one’s diet for health, eating vegetables, frequency of eating breakfast, frequency of family breakfasts, using information on nutrition labels, and positive conversations with family or friends during meals were positively associated with higher household income. After adjusting for all variables, frequency of family breakfast and dinners were both negatively associated with higher household income. Frequency of eating breakfast was not associated with household income. For all income categories, all domains except frequency of eating breakfast showed significant trends.
In the unadjusted analysis (Table 4), taking care of one’s diet for health, eating vegetables, using nutrition information, and positive conversations during meals were positively associated with completing 4-year college or graduate school. Frequency of family breakfasts and dinners were negatively associated with completing 4-year college or graduate school. After adjusting for all variables, frequency of family dinners was not associated with education. For all education levels, all domains except frequency of eating breakfast and frequency of family dinners showed significant trends.
Results of the logistic regression analysis of the association between household income and eating behaviors for men are shown in Table 5. In the unadjusted analysis, taking care of one’s diet for health, eating vegetables, frequency of eating breakfast, frequency of family breakfasts, using information on nutrition labels, and having conversations with family or friends during meals were positively associated with higher household income. After adjusting for all variables, frequency of family breakfasts and dinners were negatively associated with higher household income. The AOR of household income for frequency of family breakfasts and the 95 % CI calculated by residence status among men were: middle-income households with men living with other people: AOR 0.68, 95 % CI 0.48–0.97, p = 0.032; high-income households with men living with other people: AOR 0.70, 95 % CI 0.48—1.02, p = 0.065; and middle-income households with men living alone: AOR 0.82, 95 % CI 0.26–2.60, p = 0.730. It was impossible to estimate an AOR or 95 % CI for high-income households with men living alone because of the small sample size. Men with the highest levels of income were positively and significantly more likely than those with the lowest income levels to take care of their diet for the sake of their health, eat vegetables, use nutrition information, and converse during meals at the recommended levels. There was no significant association between frequency of eating breakfast and household income, but positive associations were seen between men’s household income, and the domains of taking care of one’s diet for health, eating vegetables, using nutrition information, conversations during meals, and negative association was seen between frequency of family breakfasts.
In the unadjusted analysis for women (Table 6), most eating behaviors were positively associated with higher household income. After adjusting for all variables, frequency of family dinners was negatively associated with the highest household incomes. The AOR of household income for frequency of family dinners and the 95 % CI by residence status among women were: middle-income households with women living with others: AOR 0.83, 95 % CI 0.62–1.12, p = 0.225; high-income households with women living with others: AOR 0.65, 95 % CI 0.47–0.90, p = 0.010; and middle-income households with women living alone: AOR 1.90, 95 % CI 0.51–6.99, p = 0.337. The AOR and 95 % CI were impossible to estimate for high-income households with women living with others because of the small sample size. Women with the highest levels of income were positively and significantly more likely than those with the lowest levels of income to take care of their diet for the sake of their health, eat vegetables, use information on nutrition labels, and engage in conversation with family or friends during meals at the recommended levels. Frequency of eating breakfast and frequency of family breakfasts were not associated with household income. For household income among women, all domains showed significant trends.
Results of the logistic regression analysis of the association between education and eating behaviors for men are shown in Table 7. In the unadjusted analysis, taking care of one’s diet for health, eating vegetables, using nutrition information, and conversing with family or friends during meals were positively associated with attainment of a 4-year college or graduate degree. Frequency of family dinners was negatively associated with completing 4-year college or graduate school. After adjusting for all variables, men who had graduated from 4-year college or graduate school were positively and significantly more likely than junior high school or high school graduates to take care of their diet for the sake of their health, eat vegetables, use the information on nutrition labels, and have positive conversations during meals at the recommended levels. Frequency of eating breakfast was not significantly associated with education. For education among men, the domains of taking care of one’s diet for health, eating vegetables, frequency of family dinners, using information on nutrition labels, and conversations during meals showed significant trends.
In the unadjusted analysis for women in Table 8, taking care of one’s diet for health, eating vegetables, using nutrition label information, and having conversations with family or friends while eating were positively associated with graduating from 4-year college or graduate school. After adjusting for all variables, women with 4-year college or graduate degrees were positively and significantly more likely than those with junior high or high school level education to take care of their diet for the sake of their health, eat vegetables, have frequent family breakfasts, use nutrition information, and converse during meals at the recommended levels. Frequency of eating breakfast was not associated with education. For education among women, the domains of taking care of one’s diet for health, eating vegetables, frequency of family breakfasts, using nutrition information, and positive conversations during meals showed significant trends.
In this study using an Internet-based survey, eating vegetables, using the information on nutrition labels, and engaging in positive conversations with family or friends during meals were positively associated with higher household incomes and education levels among both men and women. Lower frequency of family breakfasts and dinners were associated with higher household income among men; lower frequency of family dinners was associated with higher household income among women. These associations were not seen with education. We found that eating behaviors differed by SES, suggesting that supporting healthy eating behaviors could reduce health disparities. To our knowledge, this is the first study to investigate the association of SES with healthy eating behavior in Japanese adults.
Our results showed that eating vegetables is a dietary behavior that is affected by SES. We set five small dishes of vegetables or about 350 g per day as a standard against which to assess frequency of vegetable intake. This is in line with a previous study in Japan , which suggested that the number of vegetable dishes consumed might be a simple and valid measure of vegetable intake and set an intake of 350 g per day as a standard. Those results predicted higher vegetable intake among those with higher SES and healthier eating habits, and may explain the association between household income and vegetable intake found in the NHNS . In both Western countries and Japan, individuals with lower SES have a lower intake of healthy foods like vegetables [14, 16–18, 36, 37]. Many socioepidemiological studies in Japan have noted associations between SES and cancer risk  as well as cardiovascular disease and its risk factors [6, 36]. Increased vegetable intake is effective in preventing lifestyle-related cardiovascular disease and cancer . At the behavioral level, we found that individuals with lower household incomes and education levels tended to eat vegetables less often. This highlights an urgent need for dietary intervention programs aimed at people with low SES, to promote vegetable-eating as a way to lower the risks of cancer and cardiovascular disease.
We found that participants with higher SES used the information on nutrition labels. Sinclair  reported that participants with higher household income who had attained medium to high education levels were significantly more likely to answer questions correctly on calorie intake than participants with lower education levels. In other reports, people who used the information on nutrition labels had healthier diets  and higher Healthy Eating Index scores . We found that 46.3 % of Japanese adults used nutrition information, which is in line with results of the 2000 NHNS . Our study results are the first of which we are aware to report an association between using information on nutrition labels and SES in Japan. A longitudinal study that includes SES factors is needed to examine the health effects of nutrition information use.
In our sample of Japanese adults, few with a household income of 7,000,000 JPY or more reported frequent family meals. This is a new implication in this area of study. Frequent family meals in adolescence and young adulthood may have a lasting positive influence on dietary quality and meal patterns, such as greater intake of green, yellow, and other vegetables and fruit . Larson suggested that higher parental education level was positively associated with higher frequency of family meals among middle and high school students , and our findings are inconsistent with this. We found an initial positive association between household income and frequency of meals, but this became negative after adjustment. We added adjustment variables one by one and examined the results in detail, and found an effect attributable to residence status. A previous study found a high rate of skipping breakfast (15.1 %)  among men with the highest household income levels. In our study, we found a higher rate of skipping breakfast (37.2 %) in this group. We estimate that this is likely to be associated with a lower frequency of families eating breakfast together. Working women may also be unable to be at home for mealtimes because of the long working hours common in Japan (Table 5). About 25.6 % of employed men and women in Japan work are more than 49 h per week . Japan is second only to South Korea among developed countries (high-income countries defined by the Organization for Economic Co-operation and Development) for its working time ratios [41, 42]. Kuroda reported that members of higher income households had longer working hours  than those with lower incomes. Woman may work even longer hours than men because of time spent doing housework . The association between household income and the frequency of family meals did not change even by residence status, although we expected that having a family would lead to more talking at the dining table. For men with the highest household incomes living alone, the sample size was small and so could not be analyzed. We cannot therefore comment on the influence of residence status, the effect of adjusting for household income or relationships with the frequency of family meals. Further research with a larger sample size is necessary to address these issues. There are very few reports about the frequency of family meals among adults . Although we think that household income and the frequency of family meals may be related to long working hours, we did not investigate these factors over the participants’ working lifetime. Further research is needed to determine whether this result is a specific social problem associated with the longer working hours seen in Japan.
Despite participants with higher household incomes showing lower frequency of family meals, these participants conversed more about food during mealtimes. A previous study showed that Japanese children who engaged in mealtime conversations during meals had better dietary attitudes, eating behaviors, and quality of life; good health status; higher vegetable intake; and good table manners [29, 30, 46]. The results of our study help to highlight the importance of positive conversation at the dining table, which may be influenced by working parents being unable to be at home for mealtimes.
This study had several limitations. First, our sample may not have been representative of the general population because we relied on an Internet-based survey. Caution should therefore be used in generalizing the results. Internet-based survey respondents may be more likely to have certain characteristics, such as being younger, having higher levels of education and income, and having better access to the Internet [47–49]. The research company used in this study, however, periodically analyzes and updates its registrant database. Second, there is room for improvement in the questionnaire. It was difficult to calculate individual-level equivalent income because participants self-reported their household income, so we had to use category levels. Good dietary practices have been shown to be associated with good health . We did not, however, investigate certain confounding factors [14, 51] that are known to influence eating behaviors such as body mass index or health status, physical activity, and sleep status as control variables. We cannot therefore draw conclusions about the real impact of SES on eating behavior. Future research on SES and dietary habits is needed. Third, we chose to use seven items referenced in previous studies [20–30] and the NHNS  as healthy eating behaviors. There are, however, other possible healthy eating behaviors. Fourth, we cannot apply our results to people outside the target age ranges. For instance, future studies would be useful to compare our results to those for older adults and children or adolescents. Fifth, owing to the cross-sectional nature of the study, it is impossible to determine cause and effect between eating behaviors and household income or education. Despite these considerations, this study found statistically significant and positive associations between eating behaviors among Japanese adults, including eating vegetables, using the information on nutrition labels, conversing with family or friends during meals, and household income or education. It is important to examine the differences in eating behaviors among adults according to SES to facilitate early reductions in health disparity. A law was enacted in 2013 and poverty programs  were developed for children to reduce health disparities in Japan. The JAGES project  promoted the use of a large-scale socioepidemiological survey to prevent health disparities among older people in Japan. The associations between socioeconomic differences and eating behaviors in Japanese adults, however, have not been extensively investigated. It is important to promote the consumption of vegetables, as this has been shown to prevent the onset and progression of lifestyle-related diseases. Development of effective population-based program strategies to promote vegetable intake among people with low SES are also needed. Our next study will therefore aim to further identify the correlates  of eating behaviors by differences in SES.
In this study, eating vegetables, using information found on nutrition labels, and engaging in positive conversations with family or friends during meals were significantly and positively associated with higher household income and education level among Japanese men and women. We identified a need to support eating behaviors that can help to reduce health disparities. SES should be considered in planning initiatives of health promotion and dietary improvement in Japan.
World Health Organization
Japan Gerontological Evaluation Study
adjusted odds ratio
National Health and Nutrition Survey
Lantz PM, House JS, Lepkowski JM, Williams DR, Mero RP, Chen J. Socioeconomic factors, health behaviors, and mortality: results from a nationally representative prospective study of US adults. JAMA. 1998;279(21):1703–8.
Lahelma E, Martikainen P, Laaksonen M, Aittomäki A. Pathways between socioeconomic determinants of health. J Epidemiol Community Health. 2004;58(4):327–32.
Kagamimori S, Gaina A, Nasermoaddeli A. Socioeconomic status and health in the Japanese population. Soc Sci Med. 2009;68(12):2152–60.
Ministry of Health, Labour and Welfare. Health Japan 21 (the second term). Tokyo; 2013. http://www.mhlw.go.jp/file/06-Seisakujouhou-10900000-Kenkoukyoku/0000047330.pdf. Accessed 19 June 2015.
Marmot MG, Smith GD. Why are the Japanese living longer? BMJ. 1989;299(6715):1547–51.
Fukuda Y, Nakamura K, Takano T. Accumulation of health risk behaviours is associated with lower socioeconomic status and women’s urban residence: a multilevel analysis in Japan. BMC Public Health. 2005;5:53.
Kondo N. Socioeconomic disparities and health: impacts and pathways. J Epidemiol. 2012;22(1):2–6.
Fujino Y, Iso H, Tamakoshi A, Inaba Y, Koizumi A, Kubo T, et al. A prospective cohort study of employment status and mortality from circulatory disorders among Japanese workers. J Occup Health. 2005;47(6):510–7.
Fujino Y, Tamakoshi A, Iso H, Inaba Y, Kubo T, Ide R, et al. A nationwide cohort study of educational background and major causes of death among the elderly population in Japan. Prev Med. 2005;40(4):444–51.
Nishi N, Makino K, Fukuda H, Tatara K. Effects of socioeconomic indicators on coronary risk factors, self-rated health and psychological well-being among urban Japanese civil servants. Soc Sci Med. 2004;58(6):1159–70.
Shibuya K, Hashimoto H, Yano E. Individual income, income distribution, and self rated health in Japan: cross sectional analysis of nationally representative sample. BMJ. 2002;324(7328):16–9.
World Health Organization. A conceptual framework for action on the social determinants of health: social determinants of health discussion. Paper 2. Geneva: World Health Organization; 2010. http://www.who.int/social_determinants/corner/SDHDP2.pdf. Accessed 13 October 2015.
Robertson A. Social inequalities and the burden of food-related ill-health. Public Health Nutr. 2001;4(6a):1371–3.
Darmon N, Drewnowski A. Does social class predict diet quality? Am J Clin Nutr. 2008;87(5):1107–17.
Darmon N, Lacroix A, Muller L, Ruffieux B. Food price policies improve diet quality while increasing socioeconomic inequalities in nutrition. Int J Behav Nutr Phys Act. 2014;11:66.
Ministry of Health, Labour and Welfare. The National Health and Nutrition Survey in Japan, 2010. Tokyo; 2012:170–1. in Japanese. http://www.mhlw.go.jp/bunya/kenkou/eiyou/dl/h22-houkoku-01.pdf. Accessed 19 June 2015.
Murakami K, Miyake Y, Sasaki S, Tanaka K, Ohya Y, Hirota Y, et al. Education, but not occupation or household income, is positively related to favorable dietary intake patterns in pregnant Japanese women: the Osaka Maternal and Child Health Study. Nutr Res. 2009;29(3):164–72.
Fukuda Y, Hiyoshi A. High quality nutrient intake is associated with higher household expenditures by Japanese adults. Biosci Trends. 2012;6(4):176–82.
Ministry of Health, Labour and Welfare. Health Japan 21. Tokyo; 2000. http://www.kenkounippon21.gr.jp/index.html. Accessed 19 June 2015.
Bazzano LA, Serdula MK, Liu S. Dietary intake of fruits and vegetables and risk of cardiovascular disease. Curr Atheroscler Rep. 2003;5(6):492–9.
Takachi R, Inoue M, Ishihara J, Kurahashi N, Iwasaki M, Sasazuki S, et al. Fruit and vegetable intake and risk of total cancer and cardiovascular disease: Japan Public Health Center-based prospective study. Am J Epidemiol. 2008;167(1):59–70.
Horikawa C, Kodama S, Yachi Y, Heianza Y, Hirasawa R, Ibe Y, et al. Skipping breakfast and prevalence of overweight and obesity in Asian and Pacific regions: a meta-analysis. Prev Med. 2011;53(4–5):260–7.
Timlin MT, Pereira MA. Breakfast frequency and quality in the etiology of adult obesity and chronic diseases. Nutr Rev. 2007;65(6 Pt 1):268–81.
Larson N, MacLehose R, Fulkerson JA, Berge JM, Story M, Neumark-Sztainer D. Eating breakfast and dinner together as a family: associations with sociodemographic characteristics and implications for diet quality and weight status. J Academy Nutr Diet. 2013;113(12):1601–9.
Eisenberg ME, Olson RE, Neumark-Sztainer D, Story M, Bearinger LH. Correlations between family meals and psychosocial well-being among adolescents. Arch Pediatr Adolesc Med. 2004;158(8):792–6.
Fulkerson JA, Larson N, Horning M, Neumark-Sztainer D. A review of associations between family or shared meal frequency and dietary and weight status outcomes across the lifespan. J Nutr Educ Behav. 2014;46(1):2–19.
Pérez-Escamilla R, Haldeman L. Food label use modifies association of income with dietary quality. J Nutr Educ Behav. 2002;132(4):768–72.
Sinclair S, Hammond D, Goodman S. Sociodemographic differences in the comprehension of nutritional labels on food products. J Nutr Educ Behav. 2013;45(6):767–72.
Kishida N, Kamimura Y. Relationship of conversation during meal and health and dietary life of schoolchildren in the case of 5th and 6th graders in urban district. Hiroshima Japan J Nutr Diet. 1993;51(1):23–30. in Japanese.
Moriwaki H, Kishida N, Kamimura Y, Takeda Y, Sakuma A, Teraoka C, et al. Relationship of the health condition, daily living habits, and diet of female university students with their mealtime conversation at elementary school. J Home Econ Japan. 2007;58(6):327–36. in Japanese.
Murayama N. Substantial study on socioeconomic factors to prescribe the content of the Japanese eating habits. Research report of Health Labour Sciences Research Grant. Japan: Murayama Nobuko; 2014. in Japanese.
Japan Gerontological Evaluation Study. 2015. in Japanese. http://www.jages.net/. Accessed 19 June 2015.
Ministry of Internal Affairs and Communications. Statistics Bureau Ministry of Internal Affairs and Communications: the 2013 population census of Japan. Tokyo; 2013. in Japanese. http://www.soumu.go.jp/menu_news/s-news/01gyosei02_02000055.html. Accessed 19 June 2015.
Ministry of Health, Labour and Welfare. 2012 Comprehensive Survey of Living Conditions. Tokyo; 2012. in Japanese. http://www.mhlw.go.jp/toukei/saikin/hw/k-tyosa/k-tyosa12/. Accessed 19 June 2015.
Ozawa K, Takemi Y, Eto K, Tanaka H, Fujii H, Ishikawa M, et al. Stage of vegetable intake and self-reported number of vegetable dishes consumed: are they valid measures of vegetable intake among middle-aged adults? Japan J Nutr Diet. 2013;71(3):97–111. in Japanese.
Fukuda Y, Hiyoshi A. Associations of household expenditure and marital status with cardiovascular risk factors in Japanese adults: analysis of nationally representative surveys. J Epidemiol. 2013;23(1):21–7.
Satheannoppakao W, Aekplakorn W, Pradipasen M. Fruit and vegetable consumption and its recommended intake associated with sociodemographic factors: Thailand National Health Examination Survey III. Public Health Nutr. 2009;12(11):2192–8.
Bowman SA, Lino M, Gerrior SA, Basiotis PP. The Healthy Eating Index: 1994–1996. U.S. Department of Agriculture, Center for Nutrition Policy and Promotion. 1998, CNPP-5. http://www.cnpp.usda.gov/sites/default/files/healthy_eating_index/hei94-96report.PDF. Accessed 13 October 2015.
Ministry of Health, Labour and Welfare. The National Health and Nutrition Survey in Japan, 2000. Tokyo; Daiichi Shuppan Corporation; 2002:55–6. in Japanese
Larson NI, Neumark-Sztainer D, Hannan PJ, Story M. Family meals during adolescence are associated with higher diet quality and healthful meal patterns during young adulthood. J Am Diet Assoc. 2007;107(9):1502–10.
Statistics Japan. Comprehensive Survey of Social Living Conditions, 2011. Tokyo; 2011. in Japanese. http://www.stat.go.jp/data/shakai/mtopics/pdf/mt01.pdf. Accessed 19 June 2015.
Honkawa Y. Social fact data pictorial record, 2013. in Japanese. http://www2.ttcn.ne.jp/honkawa/3100.html. Accessed 19 June 2015.
Kuroda S. How long do Japanese work?-A comparison before and after the shorter work week. RIETI Policy Discussion Paper Series 10-P-002. Research Institute of Economy, Trade & Industry, IAA; Tokyo; 2010. in Japanese. http://www.rieti.go.jp/jp/publications/pdp/10p002.pdf. Accessed 19 June 2015.
Organization for Economic Co-operation and Development’s Better Policies for Better Lives. Balancing paid work, unpaid work and leisure. 2014. http://www.oecd.org/gender/data/balancingpaidworkunpaidworkandleisure.htm. Accessed 28 August 2015.
Sobal J, Hanson K. Family meals and body weight in US adults. Public Health Nutr. 2011;14(9):1555–62.
Eto K, Takemi Y, Nakanishi A, Adachi M. Associations of family meal frequency and voluntary communication during mealtime with dietary attitudes, dietary behaviors, and quality of life among fifth-grade students. Japan J Health Educ Promot. 2012;20(3):192–206. in Japanese.
Yasunaga H, Ide H, Imamura T, Ohe K. Medical research using Internet questionnaire in Japan. Nihon Koshu Eisei Zasshi. 2006;53(1):40–50. in Japanese.
Eysenbach G, Wyatt J. Using the Internet for surveys and health research. J Med Internet Res. 2002;4(2):e13.
Rhodes SD, Bowie DA, Hergenrather KC. Collecting behavioural data using the world wide web: considerations for researchers. J Epidemiol Community Health. 2003;57(1):68–73.
Belloc NB, Breslow L. Relationship of physical health status and health practices. Prev Med. 1972;1(3):409–21.
Giskes K, Kamphuis CB, van Lenthe FJ, Kremers S, Droomers M, Brug J. A systematic review of associations between environmental factors, energy and fat intakes among adults: is there evidence for environments that encourage obesogenic dietary intakes? Public Health Nutr. 2007;10(10):1005–17.
Green LW, Kreuter MW. Health program planning: an educational and ecological approach. 4th ed. New York: McGraw-Hill; 2005.
This research was supported by a Grant-in-Aid for Scientific Research on Innovative Areas (No. 25560357) from the Ministry of Education, Culture, Sports, Science and Technology, Japan.
The authors declare that they have no competing interests.
SN designed the study, analyzed the data, and drafted the article. TI and KH (Hata) participated in the study design, supervised, and helped draft the manuscript. MM conceived the study, participated in its design and coordination, and supervised writing the article. MT, KH (Harada), and TA participated in the study design and supervised writing the article. All authors read and approved the final manuscript.