The de-structuration of eating models in East Asia under compressed food modernity: An empirical synthesis

The de-structuration of eating models refers to a multitude of contemporary dietary changes, such as meal skipping and eating out, that diverge from ‘proper ’ eating models in given societies. This phenomenon has been studied primarily in Western societies and diagnosed as a more modest change than previously assumed by alarming social discourse. However, this view must be relativised from non-Western perspectives. De-structuration involves the weakening of dietary normative systems and the increased food anxiety, the typical symptoms of reflexive modernity. This concept is theoretically based on the paradigm of ‘plural ’ modernities, but it has been scarcely tested empirically in non-Western regions. Web-based questionnaire surveys were conducted from 2021 to 2024 in four East Asian societies that have experienced compressed modernisation. The two studies in Japan (n = 973) and Taiwan (n = 920) have already been reported elsewhere. In this article, discussion on this Japan-Taiwan comparison is further extended with new datasets in South Korea (n = 1039) and China (n = 1035), providing an empirical synthesis of eating models and their de-structuration in four East Asian societies. In contrast to Western societies, de-structuration in East Asia has been more intense than a modest change. Similarly, in Taiwan and South Korea, the degree of change has been so large that de-structuration has extended to dietary norms. In Japan, the norm – practice discrepancy has been intensified by the country ’ s gendered dietary norms. Finally, in China, there has been a time lag between dietary changes and the drastic socioeconomic re-forms since the 1980s, manifesting an embryonic form of de-structuration. These phenomena are diverse aspects of compressed food modernity, and our article contributes by providing empirical support for plural views of food modernisation.


The de-structuration of eating models
The de-structuration of eating models has been a key but contested theme in food studies.As the prefix de-implies, this term refers to a multitude of contemporary dietary changes, such as meal skipping, eating alone and the outsourcing of family meals, that diverge from some 'proper/standard' eating models in given societies.Mass media and agri-food industries have become alarmed by the alleged dramatic dietary changes.Meanwhile, public-health experts have provided scientific evidence about the negative consequences of these changes, and sociologists have diagnosed the emergence of this new social phenomenon with some caution. 1  Despite a long series of relevant discussions about the loss of traditional and family meals, it was French sociologists who first presented de-structuration as a scientific subject.Fischler (1979) observed 'the de-structuration of normative systems and the social controls that [have] traditionally governed the dietary practices and representations ' (p. 191) as the underlying cause of increasingly individualised eating in Western societies in the 1970s.To explain such a normless situation, Fischler coined the term 'gastro-anomy' by combining the sociological concept of anomy, developed by Emile Durkheim, and the etymological sense of gastronomy as the governing of one's stomach.
To advance the empirical study of de-structuration, Herpin (1988) proposed five evaluative dimensions: de-concentration (the multiplication of mealtimes, such as snacking, and the resultant simplification of meals); de-implantation (the irregularisation of mealtimes); de-synchronisation (eating alone); de-localisation (the outsourcing of family meals); and de-ritualisation (obfuscation of the distinction between ordinary and festive meals).With this framework, Herpin studied 767 French consumers and noted 'partial responses to the hypothesis of an increasing de-structuration of diets ' (p. 521).Empirical studies on the de-structuration of the French eating model have burgeoned but yielded contradictory results.Grignon (1996), who studied university students as precursors of future dietary changes, even criticised the discourse of de-structuration as mere publicity by food industries and a group of food sociologists.
Amid this controversy, Poulain (2002aPoulain ( , 2002b) ) made several important contributions.The first was to expand the previous analytical framework, which was primarily based on actual practices (Herpin, 1988), to include essential information on dietary norms and the logic of choice.This focus on eaters' subjectivity has long been debated in the sociology of food (Halbwachs, 1912;Lahlou, 1995;Fischler, 1988Fischler, , 1990;;Poulain, 2017Poulain, [2002]]) but has not attained adequate attention in the analysis of de-structuration.Through interviews and questionnaires with French users of workplace cafeterias, Poulain (2002a) observed a 'contradiction between the idea of a proper meal and actual practice ' (p. 49).In their study, 62% of users believed that a real meal comprises at least one starter, a main course with side dishes and a dessert, but only 39% behaved accordingly at dinner.
The second contribution, advanced further in subsequent works (Poulain, 2012(Poulain, , 2018(Poulain, , 2019;;Poulain et al., 2022), was to mobilise reflexive modernity theories to diagnose such a norm-practice discrepancy.Reflexive modernity (otherwise called second modernity or late modernity) refers to living with the radical speed of modernisation and, simultaneously, confronting its side effects (Beck, 1992;Giddens, 1990).Contemporary phenomena that characterise reflexive food modernity include the individualisation of eating; weakening of dietary norms; increase in food anxiety; and emergence of alternative regulatory mechanisms, such as nutritional sciences and food heritages/traditions (Poulain, 2012(Poulain, , 2018(Poulain, , 2017(Poulain, , 2002(Poulain, , 2019)).Thus, de-structuration becomes a useful entry point to analyse the forms and processes of reflexive food modernity in given societies.
The third contribution was to facilitate international dialogues on destructuration.Along with the influential paper by Fischler (1988), Poulain's (2002a) article appeared in Appetite and became a benchmark for subsequent studies on de-structuration in anglophone literature.As noted by Poulain himself, 'only comparative studies […] shed light on the subject [de-structuration] by escaping an ethnocentric and historically bound perspective ' (p. 55).Comparative studies are also required by the thesis of reflexive food modernity itself, which claims a 'plurality' of modernities, particularly in non-Western societies (Beck & Grande, 2010).
Since the 2000s, empirical studies on de-structuration have burgeoned in many regions outside France, including Belgium, the UK, the US and Nordic countries (Holm et al., 2012;Holm et al., 2016;Lund & Gronow, 2014;Mestdag, 2005;Mestdag & Glorieux, 2005;Ueda & Poulain, 2021;Warde, 2016;Warde, 2007;Yates & Warde, 2015, 2017), and Asia (Khusun et al., 2023;Kim, 2020;Mognard et al., 2023;Poulain et al., 2014Poulain et al., , 2022;;Ueda & Chiu, 2024).Most studies in Western societies have suggested that the evolution of eating models since the 1970s has been much more 'modest' than generally assumed in social discourse on de-structuration, despite their varieties in social contexts, data, methods and periods of analysis.Nevertheless, it is important to note again that these studies primarily focused on actual dietary practices, not the norm-practice discrepancy and its theoretical implication for reflexive food modernity (Poulain, 2002a;2002b).As will later be observed in the case of Japan, de-structuration can occur even with modest dietary changes by elevating dietary goals and consequentially enlarging the norm-practice discrepancy (Ueda, 2024a).
Aside from this sociological tradition, public-health and nutrition experts also have explored specific aspects of de-structuration, including meal skipping (Ballon et al., 2019;Horikawa et al., 2011;Ma et al., 2020), solitary eating (Hammons & Fiese, 2011;Dallacker & Mata, 2018), the outsourcing of family meals (Orfanos et al., 2007;Lachat et al., 2012;Gesteiro et al., 2022) and snacking (Mesas et al., 2011;Miller et al., 2013).These studies have provided scientific evidence for the health costs of contemporary dietary changes and have contributed to disseminating the de-structuration discourse through mass media and dietary instruction books.
Obesity has been one of the most interested topics in these public health studies.In East Asia, the focus of our article, the obesity rates (e.g., 4.5% among adults in Japan and 6.2% in Taiwan: Global Obesity Observatory, 2024) still remain not as high as in the West and in South-East Asia (e.g., Fournier et al., 2016).2However, public health scholars and officials in East Asia have alarmed the nations of the potential association between de-structuration, notably the prevalence of eating out, and the rise of obesity as well as of non-communicable diseases such as diabetes and heart diseases (Pan et al., 2011;Pan, 2022;Ma et al., 2020;MHLW 2015).
This situation poses a difficult question about how to balance contemporary eaters' subjective values and health consequences.In this context, an increasing number of studies have been conducted to explore people's definition of dietary well-being.While identifying crosscultural similarities in the conceptualisation of this well-being, probably due to science and policy, these studies have also called for further comparative studies, particularly on the social and psychological aspects of eating that exhibited large cultural differences (Ares et al., 2015;Ares et al., 2016;Fischler, 2015;Jaeger et al., 2022;Lappalainen et al., 1998;Rozin et al., 1999).
Given the growing body of literature that has raised diverse issues regarding de-structuration, it would be appropriate to avoid rushing to conclusions and instead gather further empirical insights.As such, a contribution from non-Western societies is essential to prevent what Poulain (2002a) self-criticised as ethnocentrism.In particular, Asia, which has experienced food modernisation in a 'compressed' manner (Chang, 2016) and created complex food social realities (Augustine & Poulain, 2018;Ehlert & Faltmann 2019;Figuié et al., 2019;Khusun et al., 2023;Mognard et al., 2023;Poulain et al., 2014Poulain et al., , 2022)), can be a useful example.Can de-structuration of eating models also be observed in Asian societies?If so, what are the forms of de-structuration, and how can they be distinguished from those in Western societies?

Compressed food modernity in East Asia
In fact, the social discourse on de-structuration has been as popular and contested in East Asia as in Western societies.For example, in Japan, contemporary dietary changes, such as skipping meals, eating alone and outsourcing family meals, have already been problematised since the 1980s, and relevant questions were included in the National Health and Nutrition Survey to track such changes.Since the 2000s, these encompassing dietary changes gradually began to be labelled as hōshoku (the de-structuration of eating), which became a key term in social debates (NHKRI, 2006).
However, such de-structuration trends in Japan have been addressed differentlymore nutritionally and less sociologically compared to Western societies.Moreover, the relevant reports have only been effective global comparisons, but, domestically, East Asian societies define it as BM I ≥ 25 kg/m 2 according to the Asia-Pacific criteria of the World Health Organisation guideline (WHO 2000).
published in Japanese and have not been intended for international comparison, discouraging global and interdisciplinary dialogues.The same situation also applies to Taiwan, South Korea and China. 3s one of the few exceptions, Ueda's (2022aUeda's ( , 2022b) ) research group employed an analytical framework that enables international comparison with relevant sociological studies (Herpin, 1988;Poulain, 2002aPoulain, , 2002b) ) and, through a web questionnaire with 973 Japanese, reported that the norm-practice discrepancy was similarly observed in contemporary Japan.
Subsequently, Ueda (2024a) performed a longitudinal analysis of the evolution of the Japanese eating model using a variety of national statistics. 4Despite non-negligible dietary changes, Ueda (2024a) observed that the norm-practice gap has intensified since the 1980s by reinforcing dietary norms.For example, the highly demanding meal structure ichi-ju-sansai (rice, soup and three side dishes), which originates from tea ceremonies, began to be viewed as a proper meal in the 1980s as a counter-reaction to the 'de-housewifisation' of women (Ueda, 2022c).Ueda (2024a) then argued that this social complexity arose due to the 'modern family system' in Japan.The modern family system refers to the standardisation of the modern family that divides labour according to gender, with the husband as the wage-earner and the wife as the homemaker (Ochiai, 1989).It is within this modern family system that so-called 'proper' eating models (e.g., three meals a day, the positive values attached to family meals and conviviality) became standardised (Cwiertka, 2007;Ueda, 2022cUeda, , 2024a)).
In Japan, the modern family system was established during the 1950-1970s, but its prerequisite demographic and social conditions (such as a demographic bonus, housewives and nuclear families) no longer exist.Despite this reality, since the 1980s, the Japanese government has chosen to fixate on and reinforce this modern family system under a policy oriented towards a so-called 'Japanese-style welfare society', rather than restructuring the country's family policies (Ochiai, 2019(Ochiai, , 2023)).This policy choice has created difficult conditions, specifically the largest gender gap (WEF 2022) and the highest poverty rate among single mothers in high-income countries (OECD 2022a) and one of the lowest fertility rates in the world (MHLW 2024).Ochiai (2014Ochiai ( , 2023) ) characterised this peculiar condition in Japan as the result of 'semi-compressed modernity'.Compressed modernity is a concept originally proposed by Korean sociologist, Kyung-Sup Chang, to characterise the complex realities facing Asian societies. 5The concept of semi-compressed modernity is one example of such attempts to understand plural modernities in Asia, but it was specially devised to characterise the Japanese condition.In fact, Taiwan, South Korea and China, due to their compressed modernity, have not experienced a stable period (e.g., 20 years in Japan) of the modern family system, which helped these societies move toward a new stage of social development more quickly.
To advance the thesis of semi-/compressed food modernity, Ueda and Chiu (2024) applied the same method as the Japanese study (Ueda, 2022b) to elucidate Taiwan's eating model.They demonstrated that de-structuration was observed in Taiwan, and to a larger degree than in Western societies, which can be associated with its compressed food modernity.However, as the authors noted, theoretical insights must be constrained in their study, because only two cases (Japan and Taiwan) were examined. 6 In South Korea, Kim (2020) used a sociological analysis of time-use data between 1999 and 2014, revealing that family meals were being rapidly replaced by solitary meals but refuting the thesis of the temporal de-structuration of the country's eating model.However, the latter view requires further consideration if one integrates other data sources.
According to the Korean National Health and Nutrition Examination Survey, the rate of meal skipping (on the surveyed day) increased from 21.6% in 2011 to 34.6% in 2020.Using the same data source (2013)(2014)(2015)(2016)(2017)(2018), the population with the habit of meal skipping ('less than five meals per week') amounted to 41.3% for breakfast (62.1% for the young), 7.3% for lunch and 10.4% for dinner (Kim et al., 2023;Kong et al., 2019).Although it refers to the situation two decades before, studies on the working population in Seoul identified 67.3% as meal skippers ('less than three meals per day'), with time constraint (46.5%) as the main reason (Choi et al., 2003).Notably, South Korea is one of the countries with the longest working hours (OECD 2022b).A similar study reported an interesting discrepancy, where 88.8% believed in the importance of having breakfast, but 66.8% failed to do so (Shin et al., 2002).In addition to the high prevalence of meal skipping, numerous studies have also suggested dramatic changes in other aspects of the South Korean eating model. 7 In contrast to these three countries, it is difficult to grasp the whole picture of the de-structuration of China's eating model.It is probable that the de-structuration discourse has yet to be explicitly formed, mainly due to the highly unique modernisation process of post-war China such as socialist society, reform and the opening-up policy for economic development and the recent 'housewifisation' of women (Li, 2017;Ochiai, 2023).Nevertheless, the de-structuration of the country's eating model has recently received more attention.Meal skipping and the rate of breakfast skipping, which has continued to be high for the young (about 30% of this age group), has also grown in the middle-aged and elderly (e.g., 4-5 times increase for the men in their 50-60s).Concerning family-meal outsourcing, the prevalence of eating out has not changed, but the use of home replacement meals has doubled from 7 to 14% of total food expenditure.Delayed dinner time (8-9 pm), which had been viewed as a new trend associated with long working hours under post-war economic development (1950-70s), has become established.The prevalence of 'quick eating' is another dimension of this workaholic environment.Meal duration has become shorter by 5-10 min per day among working populations, with increasing age-based inequality (e.g., the elderly spend 30-40 min more per day at tables than the young).
5 Compressed modernity can be defined as 'a civilisational condition in which economic, political, social and/or cultural changes occur in an extremely condensed manner in respect to both time and space, and in which the dynamic coexistence of mutually disparate historical and social elements leads to the construction and reconstruction of a highly complex and fluid social system' (Chang, 2016, p. 33). 6The evolution of the Taiwanese eating model can be observed by national statistics.In particular, the prevalence of eating out is noteworthy.The Family Income and Expenditure Records Survey in Taipei (1981-2019) reveals that, in terms of expenditure, eating out rose from 5% of the total food expenditure in 1980 to 47% in 2019.The recent Nutrition and Health Survey (2017-2020) reports that the working population (aged 19-44 years old) who 'eat out five times or more per week' amounted to 56% for breakfast, 60% for lunch and 31% for dinner (Pan, 2022).Consequently, the Taiwanese population has faced a drastic increase in obesity, diabetes and hyperlipemia (Pan et al., 2011).  Eating out ('one or more meals per day') increased from 21.4% of the population in 1990 to 46.6% in 2014 (Kwon & Ju, 2014).Eating alone ('one or more meals per day') was observed for 55% of the employed population from 2013 to 2016, while the prevalence was high among single-person households, irregular employees and the young (Kong et al., 2018;Jo et al., 2019;Lee et al., 2020).The stigmatisation of solitary eating in relation to collectivist values in traditional South Korean society has also been contested (Cho et al., 2015;Kim, 2020).
eating out have become the targets of social evaluation, although the degrees of change are relatively limited to other East Asian societies.8

Research scope
In any case, most studies on de-structuration have been limited to domestic surveys, and inconsistencies in both the used data and definitions of each dietary trend make effective international comparisons difficult.As an exception, Mognard et al. (2023) conducted a comparative study on the eating models in several Asian countries, including three East Asian regions (Japan, China and Hong Kong).However, their study was derived from previous research with a particular focus on eating out in Southeast Asia (Khusun et al., 2023;Poulain et al., 2014).As far as the scope of our article is concerned, some limitations remain, such as the lack of norm-related data, the non-inclusion of South Korea and Taiwan and the absence of in-depth consideration about the social context of each East Asian society.
Aside from the lack of effective international comparisons, we also confirm that only a few studies have explored the norm-practice gaps in eating models, despite Poulain's initial proposal.Sociologically speaking, the norm-practice discrepancy reveals important information about de-structuration and serves as a barometer for reflexive food modernity (Poulain, 2002a(Poulain, , 2002b(Poulain, , 2012(Poulain, , 2018(Poulain, , 2019;;Poulain et al., 2022).Moreover, from an ethical perspective, it is also an effective barometer of dietary well-being.That is, a large norm-practice gap represents a decrease in dietary well-being, signifying a situation in which one fails to achieve one's valued goals (Sen, 1985;Ueda, 2022aUeda, , 2022b)).
Against this backdrop, the purpose of our article is to elucidate the eating models and their forms of de-structuration in four East Asian countries through web-questionnaire surveys using the same method.The scientific originality of this article consists of three elements: First, following the initial proposal by Poulain (2002a), we utilise insights both about norms and practices to fully recognise eaters' subjectivity.Second, while the major results on Japan (Ueda, 2022a;2022b) and Taiwan (Ueda & Chiu, 2024), have already been reported, the new results on South Korea and China are added to enable a more complete comparison.Third, through this empirical synthesis, we also propose an alternative view concerning de-structuration and reflexive food modernity to the previous Western-centric ones.

Method
The method is fundamentally the same as in previous studies (see Ueda, 2022aUeda, , 2022b;;Ueda & Chiu, 2024, for full descriptions), so here we limit ourselves to describing essential information about the survey protocol.

Questionnaire development
Following previous literature on the sociology of food (Herpin 1989;Poulain, 2002aPoulain, , 2002b;;Ueda & Poulain, 2021), the questionnaire covered nine dimensions of eating models: meal frequency, place of eating, timing of meals, meal duration, persons to eat with, meal content, place of procurement, quality of food and pleasure of eating.The first seven dimensions, which are highly relevant to the de-structuration literature, were analysed for this paper.Place of procurement is not usually included in the de-structuration literature, but we consider its importance as an effective indicator of food access and food systems in each society.Some questions (e.g., meal content) were asked separately by breakfast, lunch and dinner.In the targeted four countries, breakfast is generally believed to be the first meal of the day eaten in the morning, while lunch as the second one eaten around noon, whereas dinner as the third one eaten in the evening.Our questionnaire simply listed closed questions containing the terms such as 'breakfast' without such defining descriptions, and it was ultimately each respondent who defined which meal was breakfast.
Each dimension was explored in terms of dietary norms and practices.Here, 'norms' signify a person's desired/valued level of achievement, whereas 'practices' refers to their actual achievement (such as the previous day and the previous week).These achievement levels were determined by asking questions such as 'At what time would you like to start dinner?' (norms) and 'At what time did you start dinner yesterday?' (practices).

Translation and modification
The translation of the questionnaire was conducted as follows.(i) The Japanese version was used as the source version, and the translated versions (Korean, Traditional Chinese and Mandarin Chinese) were developed by bilingual researchers (including non-author collaborators).Cultural adaptations (e.g., exemplified food items, sociodemographic status) were integrated that could retain consistency with existing statistics and social contexts in the other three countries.(ii) Preliminary interviews using the provisional questionnaires were conducted with 30 Taiwanese, 15 South Korean and 20 Chinese of various backgrounds.The question items (detailed below) and wording were also modified based on the interview insights to develop the final versions.
This second step included minor modifications on two question items.First, closed questions asking about meal frequency, 'How many meals per day would you like to have?' (norms) and 'How many meals did you have yesterday?'(practices), were added in the Taiwanese, South Korean and Chinese versions.However, in the Japanese version, meal frequency was indirectly calculated from other relevant questions (such as meal content).To enable effective international comparisons, we thus referred to national statistics (if available) as well as the national single-mother survey on meal frequency using the same closed questions (Ueda, 2024b).Single mothers are a social group with one of the highest rates of meal skipping in Japan, so this data was treated as a reflection of the upper limit.Second, question items about meal content were expanded.In the Japanese survey, meal structure was calculated based on answers to the closed questions about dish components (e.g., 'rice', 'miso soup') following an existing assessment tool (Kudo et al., 2017).However, it was implied that this assessment tool may have not matched the surveys in Taiwan, South Korea and China.In these surveys, therefore, closed questions concerning meal structure (e.g., the 'staple, soup, two dishes' pattern) were constructed based on the above-mentioned preliminary interviews and added to enhance the accuracy of participants' answers.

Operational definitions of de-structuration, the Norm-Practice discrepancy
Prior to describing the results, it is useful to clarify the operational meanings of de-structuration.This term means divergence from the 'proper' eating model in a given society; however, analytically, there are two ways of conceptualising this subject, and both are used in this paper.
The first way, which most studies have followed, is to assume some 'proper' eating model a priori and collect only the practice data before estimating the degree of their divergence.In existing studies in Western societies, this proper eating model generally comprises eating three meals per day, attaching importance to home-cooked meals, following regular meal hours, eating slowly and eating with others-an assumption that is also employed for our article. 9 Another method is to simultaneously consider the dietary norms expressed by respondents that are not necessarily consistent with the aforementioned assumption (Poulain, 2002a(Poulain, , 2002b;;Ueda, 2022aUeda, , 2022b;;Ueda & Chiu, 2024;Ueda & Poulain, 2021).This analytical distinction is important because whether the dietary norms are de-structured in relation to the practices has different sociological and ethical implications.As will be seen later, this case applies to societies experiencing compressed food modernity, including South Korea and Taiwan, and the analytical framework must thus capture such a complexity of the de-structuration.

Data collection
The participants of the web-based questionnaire were recruited via an e-mail invitation letter sent by the research company, GMO Research, Inc., which has wide service coverage in East Asia.The email included an explanation of the modality of the surveys and a link to the questionnaire sheet.Once the participants agreed with the objective and content of the surveys, they provided their informed consent by clicking on the survey's front page and completed the questionnaire sheet.The surveys were conducted from 2021 to 2024 (Japan: 2021, Taiwan: 2023: South Korea: 2024, China: 2024).To avoid seasonal bias, all the surveys were completed in March.As stated in the invitation letter, the participants received incentives (roughly 1-5 dollars).The inclusion criteria were not being a student and being aged 20-69 at the time of participation.The surveys were in accordance with the Declaration of Helsinki and approved by the ethical committees of the authors' affiliated universities (blinded for review).
The respondents' profile is summarised in Table 1.When collecting the data, the gender and age groups were adjusted for national representative samples in each country, while other attributes were not adjusted.Prior to the analysis, it is beneficial to mention the resulting sample bias.
In all four countries, the respondents had relatively high education levels. 10In terms of occupation, there were fewer manual workers (including farmers) and more clerical workers than the national averages. 11The respondents' income levels generally matched the national five-class ranges, although there were slightly more low-income groups in Japan and South Korea and high-income groups in China.This sample bias may have been caused primarily by the methodological nature of the web-based questionnaire survey.Although some caution is required in generalising the results to the national trends, this cross-culturally common sample bias may not be a major obstacle for effective international comparisons.
Country-specific sample biases also deserve special mentions.Singleperson households were under-represented in the three countries Note 1 Numbers were rounded up and totals may not necessarily add up to 100%.Note 2 Income levels were grouped based on national five-class ranges.In China, separate ranges were used given its large urban-rural income disparity.Note 3 In the Japanese survey, the different method of asking about family structure (e.g., 'whether or not to live with children') was different, and it was not possible to distinguish between nuclear families, extended families and other family structures.Note 4 The classification was made based on the current place of residence, not the family register place.The definitions of urban and rural areas corresponded with the statistics of each country.Note 5 The categories 'majority' and 'minority' were used purely in terms of population percentage.In Japan and Korea, the majority refers to people other than foreigners, while in the multi-ethnic countries of Taiwan and China, it means Fujian and Han Chinese, respectively.The minority refers to the 55 minority ethnic groups and foreigners in China, while in Taiwan it represents Hakka (10.0%);Mainlanders (12.8%); indigenous people (1.2%); and others, such as new immigrants (2.2%).
9 Strictly speaking, this a priori assumption does not necessarily mean that East Asian populations have historically embraced such standardised dietary norms.Our study is restricted to analysis within a specific timeframe, 2021-2024, so it cannot investigate the historical transition of each eating model.Nonetheless, the existence of a large body of studies on de-structuration in East Asia seems to partly support the validity of such an assumption.excluding Japan. 12In China, the rural population was underrepresented, with only 6.3% in our survey, compared to 35.5% of the total population (NBSC 2022), which is largely due to the 'digital divide' between urbans and rural areas (CNNIC 2023).

Statistical analysis
Statistical analysis was conducted to investigate the following two issues.The first was to confirm the norm-practice discrepancy, whereas the second was to assess the effects of socio-demographic status on the eating models.Among various socio-demographic variables, gender and age, which were highly relevant in the Japanese survey (Ueda, 2022b), were selected for testing in our survey.
Regarding the norm-practice gap, the Wilcoxon signed-rank test (a nonparametric test for paired samples) was used for the quantitative parameters (e.g., meal duration), while McNemar's test was applied to the categorical/binary data (e.g., eating alone).
In terms of the sociodemographic effects, multiple regression analyses were performed for the quantitative dependent variables (such as meal duration: 10− 120 min), whereas binominal logistic regression analyses were performed for the qualitative ones (that can be expressed Yes: 1 or No: 0).As independent variables, gender and age were converted into five dummy variables in all the analyses: 'women', '30s', '40s', '50s' and '60s' (baselines: 'men' and '20s'), all of which were entered into the model (forced-entry method).
While we set 5% as the significance level for testing the norm-practice gap, we made a Bonferroni adjustment to 1% as the significance level for testing gender and age effects to avoid the problem of multiplicity (Tsuchiya, 2014).The R software (ver.3.1.2) was used for all the analyses.

Meal frequency
Meal frequency refers to the number of meals taken per day, in other words, 'the moment of the day' component of the temporal dimension (Poulain, 2002b, p. 58).To retain consistency with previous literature and dietary discourse on de-structuration, we use the term 'meal skipping' to refer to less than three meals per day; but, it is important to keep in mind this term's normative nuance.
Fig. 1 shows the rates of meal skipping.The common feature in all four countries was large differences between meals, among which the rate of breakfast skipping was the highest.Fig. 2 provides a more indepth understanding, including both norm and practice data.As mentioned in the method section, the Japanese survey substituted data from single mothers (Ueda, 2024b).Despite the single mothers' high rate of 28.3%, meal skipping was more prevalent in South Korea and Taiwan, with rates of 58.3% and 37.0%, respectively.
While the high meal-skipping rate in South Korea was consistent with existing literature (Choi et al., 2003;Kim et al., 2023;Kong et al., 2019), it is interesting to note that those without the 'three meals per day' norm amounted to 50.1% of the respondents.This result contrasts sharply with the previous situation, in which almost 90% of the urban population still had such a dietary norm (Shin et al., 2002).As such, the 'three meals per day' norm seems to have greatly weakened in South Korea over the past two decades.
Although not as drastic as in South Korea, the weakening of the dietary norm and the high prevalence of meal skipping was also observed in Taiwan.In contrast, as suggested by existing literature (Zhao et al., 2023;Chen, Li, & Du, 2021), meal skipping in China was not as prevalent, with a low rate of 9.3%, although a slight norm-practice gap was shown.

Place of eating
Fig. 3 shows the frequency of meal outsourcing (i.e., eating out and ready-made meals).For breakfast, two patterns were noticeable.In Taiwan and China, people frequently outsourced breakfast, while this was not the case in Japan and Korea.For lunch, a large proportion of the respondents in each country outsourced this meal twice a week or more, mostly due to their out-of-home office work.
Below, dinner is examined more closely, because this meal has been debated in the de-structuration discourse in close relation to household (mothers') meal responsibilities.Fig. 4 contains two types of information: One information (%) at the right bar refers to those who idealised no opportunity for outsourcing dinner and those who, whatever their  First, the norms related to dinner outsourcing are examined.The percentage of respondents who answered '0 times a week' (no outsourcing) was 44.7% in China, while this number remained around 30% in the other three countries, meaning that about 70% wanted to outsource dinner.However, the ideal frequency of out-of-home dinners differed (median [interquartile range]), with the highest being Taiwan (3 times per week), South Korea (2 [2, 3] times) and Japan (2 [1.5, 2.5] times).The situation is somewhat contrasted as the 'home-cooked meals' norm continued to function strongly in Japan, while it had weakened in Taiwan.
Next, the analysis of norm-practice gaps highlights the unique pattern in Japan, which had a large upward gap, meaning that the respondents wanted to outsource family dinner but failed to do so.Notably, there again remains a strong gendered norm that idealises home-cooked meals (Ueda, 2023a;2024a), which may have inhibited Japanese eaters from outsourcing dinner freely.
Last, the focus on actual practices reveals that the respondents in Taiwan had highest frequency of out-of-home dinner (4 times), followed by South Korea (2.75 times), Japan and China (1 [1, 3] times, respectively).These figures imply that the practices with high frequencies of dinner outsourcing in Taiwan and South Korea were closely related to weakened norms.In these two countries, where out-of-home dinners appear quite accessible, the norm-practice gaps were directed upward, meaning that some respondents ate out too often.
In Japan, although there was no observed norm-practice deviation, the extended frequency distribution (norms [1.5, 2.5] to practices [1, 3]) signifies some polarised realities, with some who eat out too often and others who do so too little.In China, this dilemma regarding eating out was not observed.

Timing of meals
Fig. 5 shows the starting times of each meal.In each country, three regular peaks are identified that correspond to breakfast at 7 a.m., lunch at 12 a.m. and dinner at 6− 7 p.m. Lunch was largely synchronised across countries, but there were regional differences for breakfast and dinner.As in the case of the place of eating, dinner is discussed below (Fig. 6).
Given that the peak time for dinner in each country was 6-7 p.m., those who started dinner after 7 p.m. are considered to be delayed.While this was not as common in Taiwan (14.8%) and China (8.8%), delayed eaters amounted to 24.7% in South Korea and 31.8% in Japan, where 12.5% started dinner at 9 p.m. or later.This situation should be distinguished from regions such as Spain, where dinner after 7 p.m. has traditionally been the standard (Diaz-Mendez & Callejo, 2013).In contrast, delayed dinner has been viewed in Japan as a negative consequence of long working hours (MHLW 1997;Ueda, 2024a).Although the degree of delay differs, the norm-practice gap was observed in all four countries, representing a failure to start dinner at regular hours.

Meal duration
Fig. 7 shows the actual duration of each meal, as well as the norm of dinner duration.In each country, breakfast time was the shortest, while dinner time was the longest, with roughly 10-20 min for breakfast, 20-30 min for lunch and 30 min for dinner.Considering the total meal duration per day across countries (median [interquartile range]), the longest in descending order were Taiwan (80 min), South Korea (70 [60, 90] minutes), China (70 [50, 90] minutes) and Japan (60 min).This ranking structure generally also applied to each meal, although there was some replacement in the rankings for South Korea and China.Fig. 7 also includes the rate of 'quick eaters', who spent only 10 min or less for every meal.In each country, nearly half of the respondents (41-54%) ate breakfast quickly.Although the norm data is not shown in the figure, those who idealised 'eating quickly' at breakfast were 28.8% in Japan, 31.3% in Taiwan, 30.3% in South Korea and 44.1% in China.This result implies that the respondents failed to eat slowly enough (and the magnitude of Japan's gap was particularly large).
The situation for lunch and dinner is even more intriguing.While it was hardly a problem in Taiwan and China, there were relatively more quick eaters in Japan at lunch (28.0%) and dinner (10.3%) (along with 14.2% quick lunch eaters in South Korea).Moreover, there was a relatively large norm-practice discrepancy for dinner in Japan.These findings, in addition to the aforementioned delayed dinner, imply that the temporal aspects of Japan's eating model are impoverished.

Persons to Eat with
Fig. 8 shows the rates of solitary eaters for each meal.In all countries, eating alone was the most prevalent at breakfast, while it was the least       frequent at dinner.While the high rates of eating alone at breakfast (47-63%) are notable, dinner is analysed more closely below (Fig. 9).
Regarding the commensality norms, most of the respondents (77-93%) wanted to eat with others.However, in practice, solitary eaters at dinner amounted to 25-27% in the three countries, excluding China, which had a low prevalence rate (9.8%).
Currently, our findings show that eating alone was the most prevalent in Japan, but the rates of this behaviour may be higher in the other three countries given the under-representation of single-person households in our sample.The norm-practice gaps were larger for Japan and South Korea than those for Taiwan and China, implying the relative strength of family norms in the former two countries.

Places of procurement for fresh ingredients
Fig. 10 shows the degree to which each procurement method for fresh foods was utilised, which clearly matched with the contemporary evolution of food systems in each country.The first trend concerns whether traditional markets, most of which consist of small-scale professional retailers, have disappeared.In Japan, according to the Census of Commerce, the share of food sales by professional retailers has dramatically fallen since the late 1960s, and it is currently below 20%.Instead, the share sold by supermarkets has increased to about 40%, and new types of food retailers, such as convenience stores and discount supermarkets, have also emerged in recent years (Ueda, 2024a).Our results are consistent with this trend, reporting the frequent usage of supermarkets (1.92 points) and discount supermarkets (0.79) in contrast to the unpopularity of professional retailers (0.30).
In South Korea, the mainstream usage of large grocery marts (such as Lotte Mart), as national statistics suggest,13 was supported by our results, with 1.96 points for large marts and 0.80 for supermarkets.Like Japan, procurement from professional retailers was unpopular (0.82 points) in South Korea.In contrast, despite the recent increasing presence of supermarkets, traditional markets remained one of the most predominant procurement places in Taiwan and China (1.81,2.06 points).
Another recent change concerns the emergence of online food procurement.Previous studies have indicated the rapid growth of the online food-retail market in South Korea and China, particularly since the COVID-19 pandemic (Tamura, 2021;Yan & Watanabe, 2023).Our study also indicated the steady growth of online procurement, with 0.78 points in South Korea and 0.46 in China, levels that are a several times higher than in Japan and Taiwan, although it still not being the mainstream method.143.7.Meal structure Fig. 11 shows the distribution of meal structures.If there is a focus on the 'staple and two or more dishes' pattern, lunch and dinner in all four countries tended to be more plentiful than breakfast.For breakfast, the 'staple only' pattern (e.g., rice balls, breads, steamed buns) was highly prevalent in all the countries, with high rates of 54.6% in Taiwan and 36.6% in China.This result relativises the prevailing Japanese discourse that has problematised the nutritional deficit of this meal pattern (Kudo et al., 2017).
On the other hand, for lunch and dinner, the comparative simplicity or possible scarcity of meal content in Japan is noteworthy.The 'staple and two or more dishes' pattern was the mainstream structure for the other three countries, while it remained 44.5% even for dinner in Japan (whereas the 'staple only' pattern was the mainstream for lunch).
Dinner is more closely examined using the 'staple and two or more dishes' and 'staple, soup and three or more dishes' patterns as two effective references (Fig. 12).In Japan, these two meal patterns, although generally believed to be too demanding to prepare for ordinary meals, have become normalised in the nutritional discourse since the 1980s, the exact period during which the modern family system started collapsing but being reinforced institutionally (Ueda, 2024a).
Interestingly, the practice rates of these patterns in the other three countries were higher than in Japan.The practice rate of the 'staple, Fig. 10.Procurement Method by Usage Degree: Practices.Note To the question 'Where did you buy fresh foods last month?Please choose your three most frequently used places' (3 points: the most frequently used, 2 points: the second most frequently used, 1 point: the third most frequently used, 0 points: other).The choices were culturally adapted to each country (e.g., 'discount supermarkets' for Japan, 'large marts' for South Korea).

Fig. 11. Meal Structure: Practices
Note To the questions 'What meal structure do you think represents a good breakfast/lunch/dinner? (norms) and 'What did you have for breakfast/lunch/ dinner yesterday?(practices), respondents were asked to choose predetermined meal structures.Only in the Japanese study was the meal structure calculated from dish components (e.g., rice, soup).Other meal structures (e.g., no staple) were omitted from this figure .soup and three or more dishes' pattern was merely 6.3% in Japan but rose to around 25% in Taiwan and China.Moreover, this norm-practice discrepancy, that is, the simplification of dinner, was a common feature in all countries.However, the depth of the gaps in the two meal patterns rose to more than 20%, implying strong gendered norms that deviated from the dietary reality.

Sociodemographic effects
The statistical results for Japan and Taiwan have already been reported elsewhere (Ueda, 2022b;Ueda & Chiu, 2024), so only the aggregated values and significant differences are presented in Table 2.For South Korea and China, detailed statistical results are included.Following previous studies, the selected eating behaviours (such as meal skipping and eating alone at dinner) were tested for sociodemographic influences.
In Japan, previous studies (Ueda, 2022b;2024a;2024c) have indicated larger gender-and age-based inequalities in people's abilities to achieve dietary well-being than income-based ones.These inequalities are coupled with overall societal problems, such as contradictions with the modern family system and a super-aging society.Here, gender-based inequalities refer to men's lower food-related abilities and women's larger burdens of family meal preparation, both a product of the modern family system.The results in Table 2 also highlight some aspects of the Japanese dietary condition.
Interestingly, such gender inequalities were not salient in Taiwan and China.On the other hand, certain gender disparities were observed in South Korea.Specifically, Korean women ate dinner alone less often (i.e., ate together more frequently) than the men, while women skipped meals more often and their dinners were not as rich in content as the men's.One may argue that the lack of gender disparity in dinner outsourcing indicates that women's responsibility for preparing family meals has been relaxed, but the aforementioned higher rate of eating together with family members at dinner among women seems to contradict such an optimistic view.
In terms of age disparities, the young in all countries had lower abilities to achieve dietary well-being than the elderly, although the degree differed.In South Korea, the young skipped meals (67.7%), ate alone at dinner (32.8%), outsourced meals more often (2.37 times per week), had more simplified dinner content and used traditional markets for fresh foods less often (0.63 points) than the elderly.In China, except for meal content and procurement methods, the inequality trends were similar to those in South Korea.However, it is noteworthy that in China there existed a more than two-fold gap between the young and the elderly in terms of meal skipping, eating alone and outsourcing dinner.In Taiwan, traditional markets, which characterise its food systems, were used by the young less often than the elderly.

Discussion
The central questions in this article were in what forms the destructuration of eating models occurs in East Asia and what contributions can be made from East Asia to the thinking about food modernity.This section is devoted to answering these questions.

Forms of the de-structuration of eating models in East Asia
In this paper, we aimed to understand the de-structuration of eating models from a combined perspective of dietary norms and practices.The first approach was to focus primarily on practices.In this conceptualisation, previous studies on Western societies have generally agreed that the de-structuration which has occurred since the 1980s has been more modest than suggested in social discourse (Holm et al., 2012;Holm et al., 2016;Lund & Gronow, 2014;Mestdag, 2005;Mestdag & Glorieux, 2005;Warde, 2016;Warde, 2007;Yates & Warde, 2015, 2017).
Our results show that this evaluation seems to contradict destructuration in East Asian, which involved high rates of meal skipping, meal outsourcing, irregular dinner hours, quick eating, solitary eating and simplified meal structures.There were internal differences in the degree of de-structuration in the four East Asian countries.Nevertheless, it can safely be said that de-structuration is the social reality in East Asia, and its degree of change is much more intense than 'modest'.
On the other hand, following the initial proposal by Poulain (2002a), we have emphasised the importance of integrating eaters' subjective information into diagnosing the de-structuration of eating models.From this composite angle, Japan was the most convincing example, with the largest norm-practice discrepancies in several dimensions of eating among the four East Asian societies.This situation seems similar to the case of France (Poulain, 2002a;2002b), but its unique social condition must be considered in relation to the modern family system (to be discussed later).
Moreover, in Taiwan and South Korea, de-structuration extended not only to dietary practices but also to norms.It is possible to argue that the weakening of norms may have been strongly associated with practices that deviated greatly from the standard.Although our study is unable to determine whether the former is the cause or the result of the latter, the cases in these two countries pose a methodological question about assuming some 'standards' a priori when investigating contemporary eating models.

Reflexive food modernity: Asia is not homogenous
The next task is to extract implications for reflexive food modernity theories from these empirical insights.In East Asia, dietary evolution since World War II has been characterised with various terms, such as 'Westernisation of the diet' and 'nutrition transition' (e.g., Drewnowski & Popkin, 1997), all of which, however, have been framed within the convergence view of modernisation (for its discussion, see Poulain, 2021).Faced by the increasing difficulty of explaining the complex food landscapes of each region, this traditional view of modernity has been gradually replaced by the perspectives of plural modernities, second modernity, reflexive modernity and, particularly in East Asia, compressed modernity (Beck, 1992;Beck & Grande, 2010;Chang, 2016;Giddens, 1990).
One of the contributions of our study is that it provides empirical findings on dietary models and their de-structuration in four East Asian countries using a comparable method.The major benefit of this comparison is that it indicated cultural differences and supported the idea that Asia is not homogenous.To facilitate the correction of a singular idea of East Asia, the dietary state of each targeted country is summarised in Table 3.As existing literature suggests, there is a consensus that each of these four East Asian societies are experiencing compressed modernity (Ochiai, 2014;Poulain, 2018Poulain, , 2019;;Ueda, 2024a).However, there are also internal differences in the degrees and characteristics of this compressed modernity that will be highlighted below.

Country-specific summary
Taiwan is the most representative of compressed food modernity, which has two important features.The first is that actual dietary practices have deviated significantly from what the society believes as proper eating model.In particular, eating out was highly prevalent.Its median frequency (4 times per week) exceeded half of the week, and about 30% the population relied on eating out five or more times a week, which far surpasses the 'modest' dietary changes in Western societies.
Another feature is that people's norms also change significantly.Dietary norms represent people's valued ways of dietary life, so they are intrinsically important as well-being related information (Ares et al., 2016;Ares et al., 2015;Ueda, 2022aUeda, , 2022b;;Jaeger et al., 2022).Nevertheless, norms can also 'adapt' to difficult circumstances and, due to their unstable nature, must be treated carefully as the basis of ethical judgement (Clark, 2009).
Likewise, it is probable that the weakening of home-cooking norms in Taiwan has been created by objective realities, such as the dehousewifisation of women and development of the restaurant industry, which was accelerated by the economic crisis at the end of the 1990s (Chen, 2010;Wang, 2021).In fact, our results identified an honest assessment by some Taiwanese eaters that they currently ate out more than initially desired.As proposed earlier by Fischler (1980Fischler ( , 2015) ) and recently empowered by the country's food policy (Ueda, 2023b), the de-structuration of norms can also be viewed positively as occurring amid a process of reconstructing new norms that can adapt to reflexive food modernity.
In contrast to Taiwan's compressed food modernity, Japan can be characterised as experiencing 'semi-compressed' food modernity (Ueda & Chiu, 2024).Although the dietary changes are not as drastic as those of compressed modernity, typical symptoms include dietary norms that have been reinforced by gendered discourse, resulting in a large divergence from reality, and a large gender disparity in abilities to achieve one's desired dietary goals.As suggested by family sociologies, this peculiar condition in Japan is the result of the modern family system, which has lost its prerequisite social conditions but has been fixated on by familialistic social policies since the 1980s (Ochiai, 2014).Very recently, in May 2024 the Japanese government finally made the principle of food security (FAO 2009) the core of its food policy, thus amending the Basic Law on Food, Agriculture and Rural Areas for the first time in over 20 years.In promoting new food policies, it is vital to address all forms of gender inequalities in food security (Sen, 1985;Ueda, 2024c).
South Korea has created a unique dietary situation containing both the semi-compressed features of Japan and the compressed features of Taiwan.Like Japan, South Korean people suffered from impoverished mealtimes, probably due to the country having some of the longest working hours among developed countries (OECD 2022b).Additionally, South Korea had a gender disparity in the burden of family meal preparation.On the other hand, like Taiwan, norms and practices have undergone significant changes, particularly in meal frequency and place of eating, manifesting challenges in the construction of new dietary norms.
In China, despite the turbulence of economic growth and social restructuring since the 1980s, the evolution of its eating model seems to have experienced a time lag.As suggested by its national statistics, the degree of dietary change was generally the most modest among the four countries.However, we identified some embryonic trends of destructuration in the eating model, particularly within the young population.'Embryonic' does not mean that China is also necessarily going to reach compressed food modernity in its matured form, as in Taiwan.Reflexive modernity theories refute such a singular view.Rather, contemporary China is characterised by certain unique trends, such as the rise of online food purchases and the recent 'housewifisation' of women (Ochiai, 2023).Therefore, the de-structuration is likely to take a different and unique form.In this sense, China will be a genuine site for the empirical development of 'plural' food modernities, and it is hoped that our dataset can be helpful for future comparisons.

Limitations and prospects
The first limitation of this study is the sample bias.Due to the constraints of the web-questionnaire survey, low-education groups, singleperson households and rural residents (only China) were underrepresented.As occasionally referenced in the results section, the data in this study was roughly consistent with existing statistics based on national representative samples, but due to these sample and methodological differences, a complete comparison was not possible.Additionally, in this article, we analysed only gender and age disparities.The impact of a wider variety of factors on eating models, such as life cycle and generational gaps, family structure, income and urban/rural areas, must be assessed in further studies.
There are two other limitations, or, more precisely, future prospects.The first prospect concerns food policy and its monitoring system.One of the backgrounds of this study was the methodological inconsistency within and between countries in food-related national statistics.In recent years, new statistics in East Asia have developed that can contribute to a comprehensive understanding of eating models by surpassing traditional ones (e.g., nutritional survey, time survey and expenditure survey).However, the challenge is that authorities and policy orientations are fragmented, and it is necessary to retain some consistency from the integrated perspective (e.g., the principle of food security).In recent well-being studies, various indicators have been developed (Ares et al., 2016;Ares et al., 2015;Jaeger et al., 2022;Ueda, 2024c), and it is expected that this intellectual body will be more tightly linked to de-structuration studies.
Another development direction is to gain a more complete picture of compressed food modernity in East Asia by integrating a more in-depth focus on nutritional status and consumed food groups.In the destructuration literature, the traditional focus on nutrients and food ingredients has tended to be replaced by a wider interest in the total aspects of dietary life, such as time, space and social relations.Above all, the primary meaning of food modernisation in the post-war era in East Asia has been the nutrition transition (Drewnowski & Popkin, 1997), that is, a reduction in carbohydrates and an increase in fats and proteins especially from animal sources.Recently, there has been a demand for an analysis of the protein transition that aims to overcome the convergent view of food modernisation inherent in the nutrition transition and emphasises the plurality of regional developments (Poulain, 2021).However, the protein-transition debate is still Western-centric, and, as the Japan's recent food policy report acknowledges the insufficiency of national efforts (WGSFE 2021), the universal contributions that East Asia can make to the planet are uncertain.

Conclusion
In this article, we evaluated eating models and their de-structuration in four East Asian countries using a unified web-questionnaire method.The de-structuration since the 1980s has been evaluated as a modest change in the West, but in East Asia, it has manifested some unique features.In Taiwan and South Korea, the degree of change in dietary practices has been so large that its de-structuration penetrated the dietary norms.In Japan, the norm-practice discrepancy has become large by reinforcing gendered dietary norms.In China, de-structuration of the eating model has remained at the burgeoning stage, with a time lag from its dramatic socioeconomic changes since the 1980s.All these phenomena are diverse aspects of compressed food modernity, and our article contributed to the literature by providing empirical support for such plural views of food modernisation.The identified cultural differences within East Asia thus invite a revision of the conventional convergent view of a homogenous Asia.Given the possibly higher degree of change in eating models in Southeast Asian than in East Asia (Mognard et al., 2023), further comparative studies are expected.
In so doing, interdisciplinary collaboration is also indispensable.As initially provoked by Poulain (2002a), 'comparative studies […] led by specialists within both the fields of human nutrition and socio-anthropology will shed light on the subject [de-structuration and food modernisation] by escaping an ethnocentric and historically bound perspective', only by which we can establish 'a shared vision for sociologists and nutritionists' (p.55).

Fig. 1 .
Fig. 1.Meal-Skipping Rate at Breakfast, Lunch and Dinner: Practices Note The percentage of the respondents who answered less than three meals to the question, 'How many meals did you have yesterday?'.Since the same question was unavailable in the Japanese survey, the data was substituted by the National Nutrition and Health Survey in 2019, which reported the mealskipping rate on the single surveyed day.B: breakfast, L: lunch, D: dinner.

Fig. 2 .
Fig. 2. Daily Meal-Skipping Rate: Norms and Practices.Note 1 The percentage of the respondents who answered less than three meals to the questions, 'How many meals per day would you like to have?' (norms) and 'How many meals did you have yesterday?'(practices).The Japanese data was substituted by the national single-mother survey (Ueda, 2024b).Note 2 The McNemar test was used to test the norm-practice gap.*p < 0.05.

Fig. 3 .
Fig. 3. Meal-Outsourcing Frequency: Practices Note To the questions 'Where would you like to eat?' (norms) and 'Where did you eat in the previous week?' (practices), the respondents answered their desired/actual weekly frequency of meal outsourcing (eating out and readymade meals).The practice data are described in this figure, while those who outsourced their meals 0 times per week are omitted here.Habitual meal skippers were excluded from the calculation.

Fig. 4 .
Fig. 4. Dinner Outsourcing Frequency: Norms and Practices.Note 1 The right bar and the dotted lines represent the rate of those who wished to outsource and actually outsourced dinner 0 times per week.Note 2 The left bar and the plots refer to the median frequency of dinner outsourcing among those who did so once or more per week.The boxes represent the first and third interquartile ranges.The maximum and minimum values are 1 and 7 in all countries and are thus omitted.Note 3 The Wilcoxon signed-rank test was used to test the norm-practice gap.*p < 0.05.

Fig. 5 .
Fig. 5. Starting Times of Three Meals: Practices.Note The percentage of starting times for each meal (unit: an hour) reflects answers to the questions 'At what time would you like to start having meals?' (norms) and 'At what time did you start having meals yesterday?' (practices).This figure was drawn based on the practice data.

Fig. 6 .
Fig. 6.Rate of Delayed Eaters at Dinner: Norms and Practices.Note The McNemar test was used to test the norm-practice gap.*p < 0.05.

Fig. 7 .
Fig. 7. Meal Duration and Quick Eaters: Norms and Practices.Note 1 The respondents were asked to answer to the questions 'How much time would you like to spend having meals?' (norms) and 'How much time did you spend having meals yesterday?' (practices).D+ refers to the norm of dinner duration.Note 2 The right bar and the dotted lines represent the rate of quick eaters who spent 10 min or less for each meal.Note 3 The left bar and the plots refer to the median duration of each meal.The boxes represent the first and third interquartile ranges.The maximum and minimum values were omitted.Note 4 The Wilcoxon signed-rank test was used to test the norm-practice gap.*p < 0.05.

Fig. 8 .
Fig. 8. Eating-Alone Rate by Meal: Practices.Note The percentage of the respondents who answered 'eating alone' to the questions 'Who would you like to eat with?' (norms) and 'Who did you eat with yesterday?' (practices).The practice data are described in this figure.

Fig. 9 .
Fig. 9. Eating-Alone Rate at Dinner: Norms and Practices Note The McNemar test was used to test the norm-practice gap.*p < 0.05.

Fig. 12 .
Fig. 12. Meal Structure at Dinner: Norms and Practices.Note I+ and II + refer to the norms of meal structures I (staple and two or more dishes) and II (staple, soup and three or more dishes), respectively.A McNemar's test was used to test the norm-practice gap.*p < 0.05.

Table 2
Multiple regression analysis of sociodemographic effects: Practices.

Table 3
Characteristics of compressed food modernity in East Asia.