Who’s cooking? Trends in US home food preparation by gender, education, and race/ethnicity from 2003 to 2016

Outcome measures

In this study, “home cooking” refers to the sum of reported time spent in all activities that were categorized as food and drink preparation (e.g., baking, cooking, broiling, boiling; packing lunches; heating up food), food presentation (setting the table, filling salt and pepper shakers, serving the meal), kitchen and food clean-up (e.g., clearing the table, washing dishes, storing leftovers), and other food preparation activities not otherwise specified.

Explanatory measures

Covariates included survey year, respondents’ gender, age, race/ethnicity, educational level, employment status, and marital status, and presence of one or more children under 18 years of age. Gender was self-reported as male or female. Race/ethnicity was categorized as non-Hispanic white, non-Hispanic black, Hispanic, or other. As in previous work [27], age was categorized into three groups to reflect progression from early adulthood to retirement ages: 18–29 years, 30–44 years, or 45–65 years. Education was also categorized into three groups: less than high school, high school degree or some college, or college degree or higher. Employment status was defined as: not in the labor force (retired, unemployed, or other), part-time (< 35 h/week), or full-time (≥35 h/week). Marital status was defined as either married (having a spouse or unmarried partner) or single.

Statistical analysis

Stata version 14.2 (StataCorp LP, College Station, TX) was used for all statistical analyses. All analyses were weighted to be nationally representative.

Descriptive statistics (proportions testing or t-tests) were used to examine changes in the proportion of respondents who reported home cooking and mean time spent home cooking from 2003 to 2016 (min/per capita/day), by socio-demographic characteristics.

Linear regression models were used to examine changes in the adjusted predicted proportion of respondents who home cooked and the mean per capita time spent cooking from 2003 to 2016, controlling for gender, age, race/ethnicity, education level, employment status, marital status, and presence of children <18y. Each covariate was included as a set of indicator variables to allow for non-linear associations. Due to the high prevalence of individuals who reported no cooking, a two-part model was used to estimate the adjusted mean time spent cooking. Separate models were used to test the interaction between gender, education, and year as well as gender, race/ethnicity, and year in order to examine whether changes in home cooking over time were different for men and women of by educational levels or race/ethnicities. A sensitivity analysis was conducted to examine the trends in the time spent in cooking among only those who reported cooking. Wald chunk test were used to determine the statistical significance of each interaction. The margins command in Stata was used to estimate adjusted predicted proportions who cooked or mean time spent cooking. P-values of < 0.05 were considered statistically significant.

Gender

The increase in men’s cooking found here mirrors trends in Europe, where research from four Nordic countries found increases from 1997 to 2012 in the percent of men cooking, particularly those from the working and upper classes [28]. Even in 2016, however, US males’ cooking levels were still lower than in the United Kingdom a decade earlier, when a 2005 time use survey showed 60% of men (and 85% of women) in the United Kingdom cooked over one 24-h period [29]. It is unclear what accounts for this increase in US males’ home cooking, although one possibility is that the rise in popularity of food-related media has disproportionately influenced men. For example, one study found that watching cooking programs was associated with more cooking only among men [30], though additional research has found that only 28% of adults learned to cook by watching cooking shows (with no difference by gender) [19]. Others suggest that popular celebrity chefs such as Jamie Oliver have presented cooking as a masculine activity [31], potentially making it more appealing to males. At the same time, this masculinization seems to have arisen as part of “foodie culture,” or the treatment of cooking as a form of leisure or entertainment rather than labor [31–33].

Of course, having the time, money, and skill to cook as a luxury rather than a necessity is likely only possible for the middle- or upper-class. This could explain why the current study found increases in cooking only for middle- or higher-educated men, but no change for lower educated men. Increased enjoyment from cooking for men could also have contributed to increased cooking levels, as enjoyment of cooking has been linked to more cooking [34]. Interestingly, cooking research from the UK and France shows that socio-economic factors are unrelated to time spent cooking for men [29, 35].

Despite their greater increases in home cooking over time, men still lag behind women in terms of the percent who cook and time spent cooking, suggesting that women remain the primary home food preparers in the United States. The reason for the small increases in women’s home cooking, which mark a reversal of previous trends, and continued relatively high levels of cooking is not entirely clear. One possibility is that women — or at least some women — have more time available for cooking due to small declines in time spent working. Women’s labor force participation, which increased in the United States during the second half of the twentieth century, has actually fallen by about 3.5 percentage points since 2000 [36]. Less time spent in the labor force could increase time available for home cooking, which has been well-documented as a major barrier to home food preparation [11, 37, 38]. Plateaus or increases in women’s cooking time could also reflect fewer new advances in time-saving technology in the kitchen (e.g., microwaves, food processors, dishwasher) over recent years compared to the late twentieth century. In addition, strong social norms likely persist around gender and cooking: women and girls are more likely to be involved in cooking, feel confident in cooking, and pass down cooking skills to children [9]. Additional evidence shows that cooking skills and mealtime practices in general also tend to be transferred from mothers to daughters [39, 40], further propagating this norm.

Education level

There was substantial heterogeneity in cooking trends by education level. This was especially pronounced for men: while the percent of men with high school/some college and college degrees who cooked increased to 47.0% and 51.9%, respectively, the percent of men with less than a high school degree who cooked remained low (about a third). A similar trend was observed in time spent cooking, with more educated men increasing cooking time from 2003 to 2016, while those with less than a high school degree decreasing cooking time. More cooking among higher-educated men is consistent with evidence showing that higher education is associated with more egalitarian ideas about gender roles, including more equitable distributions of household labor [41]. Among men with less education, lower cooking prevalence and time spent cooking suggests greater reliance on away-from-home foods such as fast food or restaurant food, or more frequent use of foods that are faster to prepare, such as ready-to-heat and ready-to-eat convenience foods. This could be problematic for diet quality and health, as highly processed convenience foods tend to be energy-dense and contain higher levels of added sugar, saturated fat, and sodium [42–44], whereas more time spent on cooking is associated with higher intakes of beneficial foods such as vegetables, salads, and fruits [45].

Among women, this difference was reversed: women with less than a high school education were more likely to cook and cooked for longer than those with higher levels of education. This contrasts with findings from a previous study using the National Health and Nutrition Survey, which found no association between education and likelihood of being the main meal planner or preparer [46]. That study, however, asked respondents only about their status as the “main” meal preparer and not about their likelihood to cook or the amount of time they spent cooking. Interestingly, the present study found no increase in either measure of cooking for low-educated women from 2003 to 2016. This is somewhat surprising, given that economic changes such as the Great Recession, increases in food prices, and wage stagnation might suggest greater increases in cooking as people trade time to save money. These findings are consistent, however, with those in a previous study, which found that low socio-economic status households (as measured by the poverty rate) did not alter cooking patterns in response to the Great Recession [27].

It is worth noting that more cooking does not necessarily equate to more healthful cooking. Some studies have shown that low-socio-economic households have lower levels of confidence in cooking from scratch or cooking with vegetables [13, 47], and may be likely to rely on ready-to-eat meals or frozen convenience foods or fried foods. Research has shown that people in low-income/low-access neighborhoods have noted food affordability — particularly for fresh produce and other basic ingredients — as major barrier to buying and preparing healthier foods [34]. In fact, one recent study found that women who spent more time preparing meals actually had greater risk of metabolic syndrome [48]. While the current work describes trends in home cooking, more research is needed to link these shifts in dietary behaviors to changes in dietary intake and downstream effects on obesity and cardio-metabolic risk.

Race/ethnicity

There was a high level of heterogeneity in home cooking by race/ethnicity, as well, particularly for men. While non-Hispanic other men had the greatest increase in percent who cooked, a similarly high percent of white men cooked in 2016, followed by Hispanic men, and these trends were consistent for amount of time spent cooking. Non-Hispanic black men were the only race/ethnic group that did not increase cooking time from 2003 to 2016, and in 2016 spent the least time cooking of any group. For women, on the other hand, both the relative change over time as well as absolute differences between race/ethnic groups were much smaller with regards to the percent cooking, though again, non-Hispanic blacks had the lowest levels. However, non-Hispanic white women spent the least amount of time cooking in 2016, followed closely by black women, and Hispanic and non-Hispanic other women spent substantially more time cooking (> 20 min/day compared to non-Hispanic whites). The low level of cooking among non-Hispanic black men and women as well as higher levels of cooking among Hispanics is consistent with previous findings [49].

One question for future study is how amount of cooking interacts with ingredients and methods used to influence dietary intake and downstream health effects, as there is likely great heterogeneity in these aspects of cooking behavior as well. For example, previous research has shown that ethnicity and culture influence the ingredients used, such as the use of fresh or frozen food vs. more processed food and canned goods [50]. Other research has shown that the non-Hispanic black families may be more likely to use high levels of sugar, salt, and fat as well as less-healthy cooking methods like frying [51–53]. Given this study’s findings on relatively low levels of cooking among black men and women, more research into home food preparation in black households may be useful to understand how cooking (or lack thereof) might contribute to diet-related health disparities for black Americans.

Limitations

This study has several limitations. First, it was not possible to examine changes in home cooking by income level due to the high level of missingness on this variable in the data. Educational attainment serves as a useful proxy for socio-economic status, has been validated as a predictor of cardiovascular risk [54], and may relate to a household’s food and nutrition-related knowledge and skill. Income plays a related but distinct role, as financial resources — or lack thereof — can limit the types of foods households can buy, whether single ingredients to cook from scratch, processed prepared foods, or foods purchased and eaten away from home. For example, research has shown that energy-dense diets high in refined grains, added sugar, and added fat cost less than fresh fruits and vegetables, meats, and fish [55], suggesting that the cost of basic ingredients (and thus home cooking) might pose an important barrier to lower-income households beyond other education-related barriers. Secondly, the time-use data is limited only to activities reported by a single individual on a single day. Thus, just because a respondent does not spend time themselves in home food preparation does not necessarily mean they are not consuming home-prepared food, either as leftovers or prepared for them by someone else. It is unclear whether dietary value differs for consuming self-prepared foods vs. consuming home-prepared food made by someone else.

One strength of the current study is that participants report all the activities in which they participated, which are then coded as “home food preparation” (or another activity) by ATUS coders. This coding structure avoids potential problems associated with variation in perceptions of what counts as home cooking by gender, education, or race/ethnicity (i.e., what one person counts as cooking, another person might not). However, the measure of home cooking used in this study does represent multiple aspects of the cooking process, from setting the table through cooking and clean-up, as well as varying levels of cooking, from simply microwaving a ready-to-eat snack to assembling several pre-prepared items to preparing an entire meal from scratch. While time spent cooking can be considered a proxy for level of cooking effort (with more time likely reflecting a more intensive preparation process or greater likelihood of being “from scratch”), it is not possible to examine this level of detail in the current study. Future research will be needed to understand which aspects of the cooking process and what level of cooking are important for better diet and health outcomes.