Household Expenditure on Leisure: a Comparative Study of Italian Households with Children from Y- and Z-Generation

  • Simona Diliberto
  • Michele Tumminello
  • Fabio M. Lo VerdeEmail author
Original Paper


The intrinsic complexity of post-materialist society makes it challenging to investigate the connection between social changes and generations. However, the study of consumption might help in the analysis of such a connection. In this paper, we analyse empirical data of consumption on leisure of Italian households, and focus on families at a very precise stage of family life-cycle, that is, couples with teenager children. We look at consumption of households at different points in time, 2001, 2007, and 2012, in order to investigate the impact of both social change and generation of children–Y-generation in 2001 and 2007, and Z-generation in 2012–on the leisure expenditure patterns of families. Specifically, we consider secondary data of yearly expenditure on a wide range of different leisure activities, and use hierarchical clustering and logistic regression to highlight specificities in family consumption patterns on leisure, depending on both the generation of offspring and the inter-time between sampled cohorts, 2001–2012 and 2007–2012. Our analysis indicates the presence of differences between the consumption patterns on leisure of families with Y-generation children and families with Z-generation children. However, our results also point out that such differences cannot be explained by solely invoking the different generation of offspring, and that social changes should also be taken into account.


Leisure Generations Family with adolescents Household consumption Classification Multivariate analysis 

1 Introduction

The complexity achieved by contemporary societies in both the so-called Global North, and, although in very different conditions, the Global South, makes every kind of sociological analysis of social change very difficult. In fact, even recently, there have been many sociological approaches, among the most varied paradigms, that have attempted to explain the dynamics of social change and find a unitary interpretive framework, at least for the systems of societies that show comparable socio-economic and socio-cultural characteristics (Noble 2000; Vago 2003; Weinstein 2010).

Recently, studying social change with the meaning of “a change in human interactions and interrelations”, scholars felt that the approach of the sociology of the generations could have greater heuristic capacity than other theories. Although sociology of the generations is a theoretical approach that has its roots already in the works of Mannheim (1928; tr. En. 1952), and it has been invoked in many studies over time (Eisenstadt 1956), only recently it has taken on new epistemological vigour in sociological research (Gilleard 2004).

According to many scholars, the generationalist approach provides a very effective theoretical framework to systematically study social changes. Indeed, within such a framework, it is possible to analyse various aspects of social change by comparing the cultural reference system and the training model of the opinions of the different generations that live in the same historical epoch and share the same representation of significant historical events. In this regard, it is necessary to make a distinction between the terms cohort and generation. The term cohort refers to a group of individuals who were born in the same year and share the same age. Although they live in the same historical period and are exposed to the same social, political and economic changes, this is not enough to define them as a generation (Bagnasco et al. 1997). Indeed, as Mannheim argues (1928; tr. En. 1952), a generation is composed by people born in the same period who experience strong social modifications (as revolutions, wars, cultural and economic changes etc.) and are affected by them to the point of breaking with previous habits, styles of life and values transmitted from the main socialisation agents (as the family and the school). Therefore, for a cohort can be described as a generation, it is necessary that the social modifications or historical events occurred have a strong impact on people, pushing them to adopt new attitudes and behaviours that are different from those of their parents and, at the same time, are shared by the members of the same generation. In light of this, a cohort of people born in the same period of time cannot automatically be considered as a generation, in spite of the amount of social changes they experienced, unless such changes deeply influenced people lifestyles, attitudes and views of the world. Such a consideration allows one to assume that, under a generational perspective, people belonging to the same generation tend to adopt specific styles of consumption that characterise them with respect to the others.

Therefore, belonging to the same generation is determined by such a sharing of evaluation of historical events and cultural representations, (Eyerman and Turner 1998; Edmunds and Turner 2002a; Edmunds and Turner 2002b; White 2013). Moreover, studies on consumption have been of great importance in the study of social change, as consumption is a way in which individuals and groups construct their identity and representation and the meanings of social world in contemporary society. In particular, scientific research focused on differences in consumption in terms of “generational differences”. Recently, the analysis of consumption trends (in a generational key) has been carried out with the aim of identifying specific life choices that define and differentiate contemporary consumption with respect to that practiced by previous generations (Kahle 1996). Some scholars argue that the “post-materialist values” ​​and their patterns of consumption are linked to age classes and to specific generations (Corsten 1999; Rawlins 2006), although little research has been done on this topic. Consumers, generally choose different products and services during the course of their lives, and such preferences–whether it is clothing, furniture, or leisure activities–change significantly over time and from a generation to another (Leventhal 1997; Solomon et al. 2006).

Despite the fact that scholars have given considerable importance to the difference in the different phases of the individual and family life cycle (Solomon et al. 2006), the focus of consumer research aimed, for a long time, at developing marketing and communication strategies to attract younger cohorts (Grant 2004; Lindstrom and Seybold 2004). In any case, it is above all the notion of “post-materialist values” that scholars consider very relevant to understand changes in the styles of consumption that occurred in recent years, especially among young people. This concept is useful to interpret the slippage of consumer interest from the consumption of material goods to other forms of consumption, those with a high symbolic value. This type of value orientation places the possibility of expressing people’s identity at the centre of consumption dynamics, by focusing on the quality of life beyond economic security, which leads to a symbolic consumption rather than a functional consumption. As for Italy, a progressive orientation of Italian households towards a post-materialist consumption has been observed in the last decades. Indeed, according to a report published by the Italian General Confederation of Enterprises, Professions and Self-Employment in 2014 (Confcommercio 2014), the decrease of families’ income has led to an overall reduction of households’ expenditure in Italy. However, such a reduction has been more pronounced for goods than services: it has been observed a 1.4% decrease of the average annual expenditure on goods in the period 2008–2011, whereas, a 0.7% increase for services has been observed in the same period of time (Confcommercio 2014, p.3). Such an empirical evidence suggests that Italian consumers are becoming more oriented to spend their (reduced) income on services rather than goods, and, therefore, they seem to attribute a greater relevance to the symbolic value of life experience (such as the leisure time) than the one they associate with essential goods (such as food and clothes). A clear example of this orientation towards a symbolic appropriation of practices and non-material goods can be observed in both the increase of sport activities and the new role that such activities assume within people’s lifestyle. Starting from the eighties, sport has become widespread in Italy (although with regional differences), and the value that people confer to sport activities has radically changed since then (Lo Verde 2015). Indeed, it can be observed a shift from an idea of sport as merely intended as an agonistic activity to an idea of sport as a practice oriented to body and health care. This way, sport activities have quickly become an integral part of modern lifestyles, interpreted by consumers as a “socialisation mean” as well as an activity that improves the psycho-physical wellbeing (“individualist value”).

Obviously, this change has involved above all the generations that have appeared after Baby Boomers (born between 1945 and 1964), however, highlighting how each generation undergoes a different mix of cultural influences that produces and determines the relative consumption behaviour of people.

Furthermore, it should be noted that each generation determines and articulates its social practices within–and influenced by–social groups, one of which is the family unit. Then, it is conceivable that generational sub-groups may appear within a family, e.g., adolescents or young adults among the children, besides other sub-groups formed on the ground of a shared role, e.g., parents or children. Therefore, a family negotiation process, which brings to the decision of purchasing a certain type, brand, and quantity of a good, is also influenced by the various subcultural sets of tastes, needs and attitudes that characterise the different generations of family sub-groups. In such a scenario, it is necessary to take into account that social influences in adolescents’ styles of life are not only transmitted by the families, and consumption choices are not just reflecting the sense of belonging to a family unit, rather they depend on horizontal pressure from peers as well. The idea is that the reference group, outside of the family unit, plays a significant role in decisions of consumption of young people, and pushes them to adopt particular tastes and styles of consumption with the aim of recognising themselves as a part of the group. In line with this view, peer influence on adolescents’ behaviour is a domain of strong interest for several disciplines, and it is investigated from different perspectives, with the objective of analysing the effects of peer selection and influence on different aspects, such as, for instance, adolescent alcohol use and abuse (Mundt et al. 2012), college choice (Bhayani 2015), prosocial behaviour (Choukas-Bradley et al. 2015).

Therefore, the aim of this paper is to illustrate what the most evident differences are in the consumption dynamics of Italian families with children who belong to Y (born between 1981 and 1994) and Z (born between 1995 and 2010) generation.1

The empirical basis of the present work consists of microdata on household consumption (multi-purpose survey) collected by the Italian Statistics Institute (ISTAT). First, we tried to identify the relationship between the categories of goods purchased, more generally, for and during leisure time and the types of households based on family structure. Secondly, the relationship between quantity and types of goods consumed and the presence of young people in the age group under analysis are investigated. The attempt is to identify a model of consumption that can be considered a style as compared to the type of family described by Italian domestic aggregate models, on the basis of the theoretical perspective that classifies the types of family cohabitation, according to a typology founded on the basis of domestic aggregates, i.e., the “who lives with whom” (Laslett 1972).

2 Literature Review

Scholars (Eyerman and Turner 1998; Edmunds and Turner 2002a; Edmunds and Turner 2002b; White 2013) argue that the generationalist approach is one of the possible ways through which sociological knowledge builds analytical categories. These categories can be understood as “systems of recognition of specific social objects” that, in their definition, allow both the analysis of the defined social object–a social group, in this case–and, at the same time, as Bourdieu argues (1991), exclude other possible forms of categorisation and, therefore, of “grouping” (class, ethnic group, religious group, professional group, etc.). According to Bourdieu, what becomes decisive in sociological analysis is precisely the process of social recognition that some social objects receive as they are considered as a group. Therefore, the concept of generation itself takes two meanings: (i) an “analytical” meaning, as it allows one to distinguish and study a specific group, and (ii) an “ontological” meaning, since people belonging to the same generation recognise themselves as “belonging to a group” with shared attitudes and social practices. In summary, the generationalist approach, on the one hand, allows us to analyse a part of the social phenomenon that we take into consideration for comparative purposes; on the other hand, it records the “objective”–in a sociological sense–existence of groups of people born in the same period of time, which they shared as experiences and, later, as a memory of experiences, including certain phases of historical change, and some epochal events.

Those who belong to the same generation may recognise themselves as the protagonists of a symbolic, cultural and value production, responsible for recognisable practices as a specific cultural experience shared in a particular epoch. Moreover, they may recognise themselves as people who share a collective imagination, that is, a set of values, attitudes, symbols and a representation of the world, which they consider as specific to their generation. As claimed by Mannheim (1928, tr. En. 1952), there is not necessarily an “interchange” or a “group interaction” among the members of the same generation, since it is a category that encompasses individuals who do not necessarily share membership in a social circle. But there is the belonging to a shared imaginary that generates a real “complicity”, as it happens for other forms of aggregation of individuals. Mannheim defined the generation as a group of individuals born in the same historical period and in the same geographical position who interact under the influence of the same social forces and the same events. Furthermore, the characterising properties of a generation are attributed to collective memory, rather than to the age of the members (Costanza et al. 2012). A collective memory derives–and it is formed by–significant political, cultural, economic events.

However, it is not possible to consider only shared experiences as elements capable of creating a generation. Only a concrete link that has “formative force” can cause external events to turn individuals into members of a generation.

What are the social connotations of consumption of Generation Y and Generation Z? And, more specifically, the consumption associated with leisure time? To highlight its characteristics, we will try to read the dynamics of leisure consumption of households with children in a generationalist way. If we consider the family unit as a budgetary unit (McDonnell 2013) composed of members belonging to different generations, as well as of different ages, we can see how, in different historical periods, the members of a family in a specific phase of family life-cycle have allocated the different resources for different consumption modes of leisure time. According to McDonnell (2013, 309), in fact, budgetary units are.

[…] relatively durable social collectives where a substantial portion of collective activity is devoted to consumption–the selection, procurement, or enjoyment of goods, services, or experiences valued for their nonpecuniary benefits. Budgetary units are composed of members with some intersubjective orientation to each other, possessing a minimal collective identity and norms for enforcing group behavior.

According to that definition, consumption-oriented budgetary units are characterised by a specific ethos, i.e., a consumption ethos, which is different from the one of organisations that operate in the market, which are driven by a profit-oriented ethos. On the contrary, the consumption ethos of a budgetary unit is oriented to the purchase of goods and services that meet the needs of group members. Such a consumption attitude is different from the individual action whose usefulness, according to classical economy, can only be calculated “individually”. In fact, in the budgetary units (McDonnell 2013, 325),

There are multiple, potentially conflicting, benefits and logics of usefulness: multiple members, with various desires for different objects, and various uses, outcomes, or functions of any given object. […] the multiple logics of usefulness in the budgetary unit are co-present but not stably ordered because they are not derivative of a single concrete master “utility”.

Therefore, since family units are groups of individuals who primarily share the resources they consume, it is necessary to understand that leisure practices change on the basis of many variables, but most of all on the basis of three essential elements: the budget constraint, the constraint determined by the phase of family life-cycle, and the influences determined by belonging to a generation that, in general, affects the consumption styles of each family member. In short, the income constraint remains a fundamental variable that influences consumption choices. But, on the other hand, it is equally evident the symbolic value assumed by the consumption practices that end up to be, in many cases, as it happens for families of the Italian middle class, strategies of social positioning (Sassatelli et al. 2008).

Between 1981 and 2010, when Generation Y and Z were born, there have been considerable socio-economic, cultural and technological changes that have influenced the differences between generational cohorts, not only in terms of values and beliefs, but also of preferences in consumption. Within the family, consumption can be considered as a negotiation activity, the result of interaction processes that concern, not only the type and quantity of goods to be consumed, but also the ways of consumption and the distribution of goods among the family members. Access to a more structured labour market, with more equal-opportunities, such as the one experienced by the generation of Baby Boomers, which removed the monopoly of earning of the “male breadwinner”, has made family more democratic. Such a change deeply influenced the decision making process within the family unit, in particular for what concerns the allocation of resources, and, therefore, the activity of choosing consumer goods, especially of unproductive goods. Furthermore, the acquisition of innovative habits and ways of life that erupted in the 1990s, thanks to the ever-increasing diffusion of digital technologies, has led to a progressive libidinization2 (Morin 1962) of the family economy. Such an attitude can be more apparent on members of the Y-generation, always poised between the traditional family values and the external influences they are exposed to, above all, in reference to the ways in which leisure time might be spent.

In the past, scholars thought that economic mobility was so high that the effects of any kind of change in income would have “dispersed” over three generations (Becker and Tomes 1986). However, recent studies (Waldkirch et al. 2004) based on longitudinal panels have shown that the intergenerational mobility of income is less fluid than Becker and Tomes (1986) imagined. The economic stratification across generations is equally pronounced if we analyse, for example, consumption data, which show that only 8% of adult children with parents belonging to the lowest quintile with regard to consumption data, moves to the highest quintile (Waldkirch et al. 2004). In other words, in that particular segment of the population, only 8% of children consume more than their fathers. Still, at the same time, the costs of access, for some modes of consumption of leisure time, decreased and so children can consume free time by carrying out practices that their fathers were not able to carry out because they were too expensive for them. Such a difference is also reinforced by the formation of modalities of use of the free time that did not exist at the time when their fathers were young. But what generates the change in the styles of leisure consumption are mainly the so-called “styles of family consumption”, which express different “family tastes” that are a function of the income, the social position of the family unit, and the generation different members of the family belong to.

The expression of different family tastes and styles of consumption would seem to emerge much more significantly in families whose life-cycle phase is characterised by the presence of adolescents, young people and young adults, that is, individuals belonging to the Y- and Z-generations (Funches et al. 2017) who consume leisure time in a way that can be interpreted in a generationalist way. In short, the adoption of a generationalist approach, even in the selection of the cohorts of households–families with adolescent children belonging to Y-generation in 2001 and 2007, and families with adolescent children belonging to Z-generation in 2012–allows us to analyse the impact of generation on the consumption patterns of families, and effectively mitigate the impact of age.

In fact, if we consider the spending behaviour of households at three different times in the period 2001–2012 (2001, 2007, and 2012), we can observe differences between families at the same stage of family-life cycle, that is, consolidated families with adolescent children, but sampled at different historical moments. Such a choice allows us to investigate the effects of both the general economic conditions (the macro-level) and the (different) generation of children (the meso-level, or the “generational” level) on consumption of leisure time. Also the way to allocate free time with respect to the total time spent in family commitments is an important variable (Rowland et al. 1986). But even at this level the different “consumption styles” become variables that are certainly determined by the age of those who consume time, by the constraints determined by the phase of the life cycle, and by the generational belonging of the different components of the family unit.

Do families with Y-generation children and families with Z-generation children show different preferences for leisure activities? Does that can be deduced from the expenditure shares of families for different types of leisure goods or services in different years? What are the reasons that determine such differences? Are they due to the education of parents, or to the number of parents who work? Is it possible that different expenditure patterns on leisure just reflect the difference between the generation of children?

3 Data Description

We analyse secondary data collected by the Italian Institute of Statistics (ISTAT) through the Survey on Household Consumption, in the years 2001, 2007 and 2012. The survey considers a stratified sample of households, which is representative of the Italian population. The survey aims at detecting the expenditure patterns of Italian households3 and their dynamics. Specifically, the ISTAT collected data about households’ yearly expenditure on a variety of goods and services—more than 280 different expenditure categories. According to the European guidelines on Classification of Individual COnsumption by Purpose (COICOP), recorded expenditure includes: food, furniture and fittings, clothing, health, transportation, leisure, education and other goods and services. In this paper, we focus on the dimension of leisure, which includes 25 expenditure categories, as detailed in Table 1.
Table 1

Expenditure categories associated with leisure time



Innovation influence


Boat, canoe, windsurf



Musical instruments



Sport equipment






Sport events subscription



Painting and dance class









Newspapers subscription



Concerts subscription



Trip abroad



Meals and accommodation abroad



Overnight accommodation abroad



Trip to Italy



Meals and accommodation in Italy



Overnight accommodation in Italy



Travel equipment



Cafe and bakery












Analog photography






Cinema, theatre, concerts



Museum and sport events


Second column provides a description of expenditure category and last column indicates whether an expenditure category is Influenced by Economic and Technological Innovation (IETI) or not (NOT IETI)

*Food Away From Home (FAFH) category includes restaurants, fast foods and all kind of meals consumed outside the home

Moreover, socio-demographic information is collected for each member of the family, including sex, age, parental relationship with the respondent, education level, job status etc.4 Here, we use socio-demographic information–age, relationship between family members, job status, and education–to group together households that present a similar composition according to the aim of the present study. Expenditure data for each family on a specific year are used here to calculate the percentage of the total expenditure on leisure that each household allocates on each expenditure category reported in Table 1. Specifically, for each family f, the total spending in euros on the 25 selected leisure time categories, TSf, is calculated, (e.g., TSf = 1000 EUR for family f). Then, the expenditure of family f on each expenditure category (e.g., 150 EUR on category “Books”) is divided by the total expenditure TSf and multiplied by 100, in order to obtain the percentage of total spending on leisure time that each family f allocated on a given good or service (e.g., a 150/1000 × 100 = 15% of total spending on books for family f). The result of the procedure is a matrix, one for each year considered in the study (2001, 2007 and 2012), where an entry represents the relative expenditure of a household on a specific expenditure category, expressed as a percentage. Concerning families, we consider households only including a couple (the parents) and children age 13–17. Such a choice allows us to compare families with Y-generation children, as sampled in 2001 and 2007, and families with exactly the same structure but with Z-generation children, as sampled in 2012. This way we avoid that possible differences between expenditure patterns may be attributed to the different age of children. Sampled families have been further stratified according to the number of working parents (1 or 25) and the highest educational level of the parents, college or higher degree (H) or a lower level (L). Table 2 reports summary statistics of the different types of families considered in the upcoming sections of the paper.
Table 2

Summary statistics of family types

Year (Generation of children)

Number of working parents

Highest educational degree of the parents

Number of sampled households

2001 (Y-generation)













2007 (Y-generation)













2012 (Z-generation)













4 Results and Discussion

4.1 Cluster Analysis

In this section, we compare the choices of consumption for leisure activities of families with the structure described in the previous section. At first, we shall focus on families with Y-generation children sampled in 2001 and families with Z-generation children sampled in 2012. According to the classification of households based on generation of children, number of working parents and highest educational degree of the parents, eight types of families shall be considered. The percentage of expenditure (share) on each one of the 25 categories detailed in Table 1, with respect to the total expenditure on goods and services associated with leisure time, has been calculated for each household. Then, the average of expenditure shares for any given expenditure category has been calculated for each family type. Such an averaging allows us to deal with the sparseness of the dataset, in order to reduce the effect of expenditure concentration on the comparison between different family types.

Table 3 shows the relative variation of the average allocation of spending in each expenditure category and for each family type from 2001 to 2012. The table shows some variations that are negative for all the family types, e.g., trip abroad, newspapers, newspapers subscription, analog photography, and meals and accommodation in Italy. Such a negative change can be due to the impact of the economic downturn on some expenditure categories, for instance, trip abroad (probably replaced by trip to Italy, as positive variations suggest, except for families with one working parent and high educational level), while others likely reflect technological changes, which, as clarified later in the paper, are quickly absorbed by households with Z generation children, which is the cohort considered in 2012. The positive change of spending observed for all family types on sport, might reflect, instead, a social change towards hedonism and culture of the body that leaded people to attribute an increased value to sport activities (Lo Verde 2015). Such a consideration is confirmed by the overall positive variation of spending on sport equipment, which is mostly apparent for families with both parents working and a high educational level. Indeed, such family type shows an increase of more than 100% of spending on sport equipment, as well as an increase of almost 200% of the expenditure on sport activities. Moreover, according to Table 3, between 2001 and 2012, a rise in interest towards more “creative” activities, such as painting and dance class, is observed. Such a positive result, together with the increase of expenditure on leisure category “other”, suggests that Z-generation children are oriented towards a more variegated leisure consumption than Y-generation teenagers. It’s also interesting to note that households with only one parent working and a low level of education (1-L) show an increment of more than 2000% of their relative expenditure on bricolage, which might be interpreted as a consequence of the expansion of IKEA on the Italian territory6 (Blackshaw 2010). Indeed, looking at the expenditure category “furniture” (not included in the set of leisure time expenditure categories, but present in our database), a significant negative slope is observed over the same period of time. Finally, Table 3 shows a positive increase of relative expenditure on FAFH and cafe and bakery categories for all family types, although such variation is slightly smaller than the others. The fact that changes of spending on some expenditure categories are either positive or negative for all family types suggests that there might be a difference in the expenditure patterns of families on leisure time that only depends on the generation of children and the decade that separates matched samples, and it is less affected by other dimensions, such as education and number of parents working.
Table 3

Relative change of the average percentage of expenditure on leisure categories between 2001 and 2012, for each family type (1-L, 2-L, 1-H, and 2-H)

Expenditure category










Cinema, theatre, concerts*















Trip to Italy





Meals and accommodation in Italy





Museum and sport events





Overnight accommodation abroad





Painting and dance class





Sport events subscription





Meals and accommodation abroad





Concerts subscription





Musical instruments















Overnight accommodation in Italy





Travel equipment





Sport equipment





Trip abroad





Boat, canoe, windsurf





Analog photography*





Newspapers subscription















Cafe and bakery*





NA indicates that in either 2001 or 2012 (or both) the expenditure of all the households of the corresponding family type is zero. Values in percentage

*Expenditure categories for witch at least 10 % of all the households in the sample display an expenditure other than zero. Such categories are, actually, those used in the logistic regression analysis

To better investigate the latter conjecture, we performed an agglomerative hierarchical cluster analysis of the described average expenditure data. Hierarchical clustering is a method of multivariate statistics suited to reduce the complexity of a dataset by detecting groups of sampled variables that show similar patterns and creating clusters more homogenous within them than between them (Anderberg 1973). Specifically, starting from the original dataset, we calculated the Euclidean distance between the objects–family types in the vertical axis of Fig. 1 and expenditure categories in the horizontal axis–and hierarchically grouped them through the Complete Linkage Clustering Algorithm (CLCA).7
Fig. 1

Cluster analysis of households’ expenditures with children belonging to Y- or Z-generation, stratified according to number of working parents and level of education. Null spending is in grey colour. Years 2001–2012

Figure 1 reports the results of cluster analysis performed both on the sampled households (vertical axis) and on the 25 categories of expenditure (horizontal axis). Family-type labels, associated with the rows of the data matrix displayed in Fig. 1, incorporate full information about the considered family types: the first letter in the label indicates the generation of children (Y or Z); the following number represents the number of working parents (1 or 2), and the last letter indicates the highest level of parents’ education (H in case of college degree or a higher educational achievement, L in case of a lower degree). We adopt such a further categorisation of families with the aim to detect if and to what extent working status and educational level affect the choices of consumption expenditures. Expenditure categories have been divided in two groups: expenditures influenced by economic and technological innovation (IETI) in the period 2001–2012 and expenditures not influenced by innovation over that decade (NOT IETI). The idea is that such a classification might help to interpret results. Indeed, some expenditure categories, such as, for instance, those related to travelling, have been greatly affected by major changes occurred in the period 2001–2012–e.g. the burst of low-cost flight companies.

Cluster analysis clearly highlights a dichotomy between the different types of families. In fact, we observe two main groups, one composed by households with Z-generation children, and one composed by households with Y-generation children. The only exception regards families with Z-generation offspring, both parents working and high educational level (Z_2_H) that merge together with all the other categories of households at the highest distance level in the dendrogram, probably due to the high values of some specific expenditure categories, such as Sport. In this respect, sport frequency expenditure category deserves further attention. As Fig. 1 reveals, although both families with Y- and Z-generation children show quite high shares for sport activities, households with Z-generation offspring tend to allocate more income in such leisure expenditure with respect to households with Y-generation offspring. A possible explanation of such a phenomenon can be related to a new emphasis in the culture of the body erupted at the end of the first decade of 2000s (Bauman 2000; Sassatelli 2010; Lo Verde 2014). In these years, sport frequency has become a leisure activity oriented to the aesthetic body care and aimed at the achievement of common standards of beauty. Such a new interest in appearance care, might imply that Z-generation adolescents go to the gym, make fitness, and play sports more than Y-generation teenagers. According to this view, it is not surprising that families with Z-generation children belonging to the high socio-economic position (two working parent and high educational level), who care even more than other family types of health, well-being and appearance, reveal higher shares of expenditure for sport activities than other types of families with Z-generation offspring.

Figure 1 also shows that there are some leisure expenditures that are present for families with Z-generation children and either null or negligible for families with Y-generation children. Half of households with children belonging to Z-generation reveal expenditures related to painting and dance class, musical instruments, and other kind of leisure that are, instead, either null or marginal for families with Y-generation teenagers. It seems that Z-generation children determine more variegated consumption patterns and heterogeneous interests than Y-generation children do, although playing sport remains the favourite leisure activity, as confirmed by high values of the corresponding shares for all kinds of families with Z-generation offspring. On the contrary, the purchase of boats, canoe and windsurfs is more typical of households with Y-generation offspring, and it is null or almost null for the others. It is conceivable that such a result could depend on a progressive loss of interest in outdoor activities among adolescents belonging to Z-generation. Indeed, the technological innovations and the increased availability of digital devices might have had a strong impact on lifestyle of young children, pushing them to spend more and more time indoor, engaged in sedentary and possibly alienating leisure activities (e.g. instant messaging, social networks, video games etc.). It’s also possible to interpret the loss of interest in the purchase of boats, canoe and windsurfs by households with Z-generation children as a result of the different socio-economic phases experienced by families with Y- and Z-generation children considered in the present study, which have been sampled in 2001 and 2012, respectively. In fact, several events occurred since 2008–the global financial crisis, the political instability, the austerity policies and the increase of unemployment rate–strongly affected the consumption expenditures and induced families to contract and even suppress expenditures on secondary and luxury goods. Despite the economic crisis and its consequences on the patterns of consumption, it is worth to note that the expenditure for eating out remains stable and proportionally high for both generations, with slightly higher expenditure shares for families with Z-generation children.

It is worth to note that a difference between families with Y- and Z-generation offspring in leisure expenditures can be observed for the purchase of analog photography and newspaper subscription. In fact, cluster analysis reveals that all families with Y-generation children devote a not negligible proportion of their income to such leisure activities, whereas, households with Z-generation offspring show lower expenditure’s shares on both analog photography and newspaper subscription. A possible explanation of such a difference lies in the digital revolution started at the beginning of 21-st century and well-established in the following years. In this period, we witnessed the appearance of new technological tools that quickly replaced those of the past: analog cameras have been replaced by digital cameras and smartphones, while printed newspapers have been replaced by blogs, online news, and TV-news h24. It is possible to suppose that the Z-generation teenagers, called “digital natives”, adapted quickly to the technological revolution and induced their parents to adopt new digital technologies. Such an explanation can be confirmed by looking at the shares on newspaper’s expenditures in Fig. 1, which are slightly smaller for families with children belonging to the Z-generation. To better highlight these findings, we set label colours of expenditures according to last column of Table 1, which distinguishes between expenditure categories either influenced or not by economic and technological innovation throughout the period 2001–2012. It is worth to note that digitalization processes had different implications depending on the type of leisure category: in some cases, digitalization processes led to the almost disappearance of the purchase of a good, e.g., the analog photography, and the subscription to newspapers in our study, while, in other cases, they modified the way in which families consume a good or use a service, but they did not affect the portion of income allocated on the corresponding expenditure category. The latter explains, for instance, why we do not observe a negative trend of expenditure shares over time, that is, between 2001 and 2012, in several expenditure categories. For example, for what concerns the rather stable expenditure on books, we suppose that in 2012 people did not reduce the purchase of such a good with respect to 2001, but they just modify their reading habits adopting new digital format and devices (e-books and e-book readers): such a modification in the style of book consumption seems not to negatively affect the quantity of income that families with Z-generation children allocate for the corresponding good; similarly for travel expenditures, reduced costs favoured an increment in the number of trips, keeping the expenditure shares stable over time. It is also worth mentioning that–for some expenditures–the role of the generation is marginal to distinguish between different types of families, whereas other dimensions come into play. This is the case, for instance, of expenditures related to meals and accommodation abroad, which are null or marginal for all families except for households with Y- and Z-generation offspring in which both parents work. Such a result suggests that, in this case, the choice of consumption is more affected by income than by the generation or the level of education of parents.

Figure 2 reports the hierarchical cluster analysis performed on the sample of households with offspring belonging to Y- and Z-generation, stratified according to the number of working parents and their highest educational level, for the years 2007–2012. Also in this comparison, families sampled in 2007 have children who belong to the Y-generation. The aim of this second analysis is to detect if there are some differences in families’ choices of expenditure for leisure activities as compared to the years 2001–2012, in order to investigate the role played by the difference in time between people belonging to different generations.
Fig. 2

Cluster analysis of households’ expenditures with children belonging to Y- or Z-generation, stratified according to number of working parents and level of education. Null spending is in grey colour. Years 2007–2012

Cluster analysis reported in Fig. 2 shows some evidences that differ from 2001 to 2012. First of all, the composition of the two largest family clusters (see vertical axis dendrogram) is quite heterogeneous with respect to children generation and level of education. In this comparison, it seems that the discriminatory variable for the classification of households is the number of parents working, which better explains clusters as compared to both the level of education of parents and the generation of children. The only exception is related to households with Y-generation children, just one working parent, and a high educational degree, which clusters together with all of the family types in which both parents have a job. This might be due to the higher purchasing power of households with only one parent working and high level of education in 2007 (pre-crisis) with respect to the power of the same households in 2012. Figure 2 shows that households with only one working parent are inclined to low cost activities, such as bricolage, painting and dance class, and gambling. Despite all, eating out is a leisure activity in which all family types allocate a significant portion of their total expenditure.

At difference with the 2001–2012 comparative analysis, results from the 2007–2012 comparison cannot reflect the generational difference of offspring. In fact, belonging to the same generation implies the sharing of events (cultural, historical, economic and political) that create a collective memory for all members, different than the collective memory of generations that experienced other events. In case of the limited time window of years 2007–2012, it is possible to suppose that adolescents belonging to Y- and Z-generation (respectively sampled in 2007 and 2012) do not have styles and choices of consumption that strongly differ, one generation with respect to the other, simply because they share similar experiences. The borderline between Y- and Z-generation children in this case is thinner and the generational impact hardly perceived. In this case, the generation does not play a primary role in consumption expenditures, whereas the number of working parents becomes more influential.

4.2 Logistic Regression

In this section, we consider a logistic model to describe the association between expenditure patterns on leisure (explanatory variables) and a binary variable which can be either the generation of children in the families (Y- or Z-generation) or the year when household data were sampled (2001 and 2012).

Logistic regression allows one to evaluate the parameters of a logistic model aimed at estimating the probability that a binary response variable takes one of two values (not necessarily numeric values or even ordered, generation Y and Z in our case) based on the values of a linear combination of a set of predictor variables (Cox 1958).

The dataset used for the regression includes data about 977 households with either Y- or Z-generation children from 2001 (Y-generation) and 2012 (Z-generation). The eight selected explanatory variables, that is, shares of family expenditure on eight expenditure categories, used in the model are such that their entries are different from zero for more than 10% of families in the dataset. To test the overall significance of the logistic regression, 100 training and test sets have been randomly and independently constructed from the whole dataset. Each training set included 80% of data and the test set the remaining 20%. Each training set has been used to perform the regression and obtain parameters’ estimates. Then the logistic function has been applied to the test set to predict if households’ children were from Y- or Z-generation, according to recorded expenditure patterns. The threshold used on the logistic function is 0.5. The resulting classification has then been compared with the actual one for the test set, by estimating the Standardised Mutual Information (SMI) (Yao 2003; Zhang and Stewart 2016) between actual and predicted classifications (average SMI 0.053), the Matthews correlation coefficient (MCC) (Matthews 1975) (average MCC = 0.22), the percentage of correctly classified families (average value equal to 64%). A p-value has been associated with mutual information by independently randomising 1000 times the actual classification of families in each one of the 100 randomly sampled test sets and by measuring the mutual information between the random classification and the predicted one. The p-value is calculated as the proportion–with respect to the overall 1000 simulations–of random replicates of the classification that showed a mutual information larger or equal to the actual one, that is the one associated with real classification of households in each specific test set. The average p-value over the 100 independent training and test sets considered in this study is 0.037, indicating that the classification obtained according to the logistic regression is statistically significant at 5%. The same level of statistical significance (<5%) has been attained by looking at MCC in place of mutual information.

Three explanatory variables show regression’s coefficients that are statistically significant, according to False Discovery Rate correction for multiple hypotheses testing, at 1%, (p-values are reported in Table 4). More specifically, the regression’s coefficients statistically significant concern the expenditures for newspapers, analog photography and gambling. All of the three variables display negative coefficients in the regression, indicating that the share of (leisure) expenditure of households on the corresponding goods was higher in 2001–Y-generation children–than in 2012–Z-generation children. Variable “gambling” requires further explanation. Indeed, it doesn’t refer to casinos, rather to nationally specific types of gambling, namely “totocalcio”, which is the official Italian gambling on soccer games, lotteries, and all the other gambling games that share the characteristic of being managed and accessible at an institutional level.
Table 4

Parameters of a logistic regression with 8 explanatory variables, and generation of households’ children as binary response variable, that is, Y- or Z-generation



Standard error



Odds ratio unit = 1%

Odds ratio unit = 10%








Café and bakery




























Analog photography














Cinema, theatre, concerts







*Statistically significant parameters at 1% after false discovery rate correction for multiple hypothesis testing

The results indicate that the main differences between families with Y-generation children in 2001 and families with Z-generation children in 2012 strongly depend on the technological and economic innovation occurred in between and on how fast adolescents belonging to Z-generation adapted to such changes. Indeed, analog cameras diffusion and associated expenditures reduced due to the appearance and subsequent diffusion of mobile phones with cameras–the digital revolution. Similarly, newspapers have been progressively replaced by news websites, as well as institutionally managed gambling games have been replaced by online gambling (also favoured by the progressive liberalisation of the gambling market that occurred in Italy between 2006 and 2011), which are both not included in the sampled expenditure variables above. Therefore, the performed empirical analysis shows us that households with teenager children experienced the digital revolution and modified their expenditure patterns accordingly. Changes on the expenditure patterns on two of the three explanatory variables, namely analog photography and newspapers, are likely a consequence of the direct influence of children on family consumption habits, as pushed by a pressure from peers. Such an influence, could also be at the root of the change in the gambling activity of parents, as children likely induced an increased usage of the internet among the parents. In other words, according to the presented analysis, the difference between Y- and Z-generation teenagers is that Y-generation teenagers and their families show pre-digital expenditure patterns on the leisure dimension in 2001, whereas, families with Z-generation teenagers in 2012 already experienced the digital revolution, and, consequently, show reduced expenditure proportions on pre-digital goods and services. Such a difference suggests that a generational interpretation of the expenditure patterns of people should take into account the social, technological, and economic changes occurred in the time that separates the generations.

Such a conclusion is supported by a logistic regression analysis performed on exactly the same dimensions for households with Y-generation teenagers in 2007 and households with Z-generation teenagers in 2012. The only significant variable in this case turns out to be Newspapers (p-value = 0.000017). However, according to the mutual information analysis already described for the case 2001–2012, the classification model for the years 2007 and 2012 is not statistically significant (mutual-information average p-value: 0.67). Although the overall logistic model is not stable enough to changes in the training and test set, the fact that p-value of variable “newspapers” is statistically significant suggests that families with Y-generation children display a legacy to pre-digital goods and services that families with Z-generation children don’t show.

One might be tempted to claim that the observed differences between families with Y-generation children, as those sampled in 2001, and families with Z-generation children, as those sampled in 2012, just depend on the aforementioned technological change occurred in that decade, and not also on the different generations of children sampled in 2001 and 2012. In line with this consideration, a legitimate question could be: if only the technological changes affected households’ consumption, should we observe the same results independently of the different generations of members of the family units? To address such a legitimate question, we have performed a logistic regression, using the same 8 expenditure’s categories discussed above, by focusing our attention only on families made of a couple without children and age between 30 and 50 years old–a range of age which is strongly overlapping with the age of parents of teenagers–in 2001 and in 2012. In this way, we investigated the impact of technological change on consumption patterns by only considering the changes occurred between 2001 and 2012, and excluding the possible influence of children’s generation. It turns out that the only variable, which is statistically significant after the FDR correction, in the case of couples without children, is, again, Newspapers (p-value 0.00006). Furthermore, the mutual information analysis using training and test sets, as previously discussed, indicates that the classification based on the logistic regression is not statistically significant (p-value 0.073) at 5%. Such a result suggests that social and technological changes are, alone, not sufficient to determine significant changes in households’ expenditure patterns for leisure activities and that the generation of children plays a non-negligible role in determining significant differences.

Finally, some of the results obtained through the logistic regression as applied to 2001–2012 data are also supported by the trend of consumption reported in Table 5, according to a survey managed by the ISTAT (2001, 2007, 2012).8 The table shows that teenager (age 15–17) readings of newspapers and magazines drop in the period 2001–2012, which is in line with the result of a statistically significant negative coefficient in the logistic regression associated with expenditure on newspapers. Similarly, the reading of books remains rather stable over the considered time window, which supports the fact that the coefficient associated with expenditure on books in the logistic regression is not statistically significant.
Table 5

Statistics about teenager (age 15–17) readings of newspapers and magazines, and readings of books. Percentage of the sample age 15–17


Teenager readings of newspapers and magazines per week

Teenager readings of books per year

1–4 times

5 or more times

1–3 books

4–11 books

12 or more books



















In summary, the generation of children might have an influence on their ability to adapt to technological advances, and, therefore, be a genuine explanation for the faster acquaintance of their families with new technologies.

5 Conclusions

The object of the present study is to explore the role of two different generations–Y and Z–of teenagers in affecting households’ consumption behaviour oriented to leisure activities. To explore such a phenomenon, we compare the expenditure on leisure activities of families with adolescents belonging to either Y- or Z- generation, in two different time windows (2001–2012 and 2007–2012), in order to take into account the interplay between the influence of generations and the influence of social, political and economic events occurred in a shorter (2007–2012) and a longer (2001–2012) period of time, and mitigate the influence of the age of children belonging to different generations.

The findings described in the paper have several implications. First of all, they improve the research on consumption choices for leisure activities from the perspective of sociology of the generations. Indeed, although the study of generational differences has its roots already in Mannheim’s works (1928; tr. En. 1952), there is still a lack of sociological research that analyses the choices of consumption in relation to the belonging to different generations and, at the same time, are not deeply influenced by the age difference of generational cohorts.

Moreover, considering households as a budgetary unit (McDonnell 2013) allows us to give an interpretation of the empirical results that takes into account the needs of all family members and their influence, more or less pivotal, on the decision-making process. In fact, one of the main findings outlined in this article indicates that households with Z-generation children have already adapted in 2012 to new technologies and modified their styles of consumption as a consequence. We are inclined to believe that such an evidence strongly depends on the consideration that parents have of their offspring’s needs, tastes and opinions. In fact, considering the household as a budgetary unit in which the consumption choices are oriented to the benefits of all the family members, the evident decline in purchase of pre-digital goods (such as analog photography and newspapers) may depend on both the children’s desire of new digital devices (e.g. digital cameras and smartphones) and the influence that adolescents have on their parents to quickly adapt to new technologies (e.g. reading news and gambling online). Furthermore, the logistic regression performed on families without children and their expenditures for leisure activities in the same time window (2001–2012) does not reveal significant differences between households of the two cohorts, which allows us to exclude that the differences in consumption patterns highlighted for households with children just reflect and depend on the social, technological and economic changes occurred in the period 2001–2012. In other words, the changes occurred in this time period have affected more the expenditure patterns on leisure activities of households with 13–17 years old children than those of families without children, likely due to two dimensions related to family composition: 1) the presence of children that, despite their generation, influence the family’s decision-making process, and 2) the key role played by generation in modifying expenditure choices, in line with technological and economic changes. From the generationalist perspective, historical changes have a stronger impact on the lifestyle of adolescents and young adults with respect to older people, since the former are more receptive and inclined to change than the latter. As Mannheim argued (1928; tr. En. 1952), youth is the stage of lifecycle in which young people start to develop attitudes, perceptions and a style of life in a way that is more and more independent of the socialisation agents typical of childhood (family and school). Such an autonomy encourages them to re-define cultural reference models (e.g. greater importance and influence is given to peers) and develop a new awareness of their social position (political views, social class, etc.). Adolescents are, therefore, more inclined to adapt to rapid changes, in contrast to adults that have settled habits, attitudes and consciousness. In this regard, the presence of youths in households assumes a key-role for orienting expenditure choices and pushing the other members to adopt new styles of consumption: as confirmed through our empirical study, adolescents between 13 and 17 years old follow the direction of historical change and accelerate the assimilation of changes by the whole family. In light of this, the main contribution of the present study is to consider teenagers, that is people of the same age, belonging to different generations. Indeed, some studies already exist that aim at investigating the generational influence on consumption patterns (Brosdahl and Carpenter 2011; Chhetri et al. 2014; Kolnhofer-Derecskei et al. 2017), but the generational effect is mixed with the effect of the different age of generations’ members in these studies.

Although the generation of children affects family consumption patterns, our results indicate that its influence cannot be separated from the one of social changes. In fact, as the empirical analysis performed on 2007–2012 data revealed, when a short time window separates Y- and Z-generation offspring, no significant differences between households’ consumption patterns are observed, and other dimensions become more influential (e.g. the number of working parents, as cluster analysis highlighted). Such findings are in line with the literature on sociology of the generations (Alwin and McCammon 2003; Bagnasco et al. 1997; White 2013), according to which generations are not just a mere chronological sequence of cohorts, and sharing the same experiences is not enough to create a generation. For a cohort to become a generation, it is necessary that the changes are internalised by people and that such assimilation creates new and shared attitudes and views of reality that could influence the style of life specific to that generation. Although people belonging to Y- and Z- generation share different collective memory, the adolescents in 2007 (Y-generation) and the adolescents in 2012 (Z-generation) act under the influence of similar social changes, and such a reduced time window does not allow them to develop a style of consumption specific of the generation they belong to.

The analysis reported in the paper indicates that, under a generationalist perspective, the observed differences in household consumption patterns on leisure activities can only be explained through the entanglement between the generation of children and the presence of major social changes that occurred in the time window that separates the considered generational cohorts (about a decade, in our case). Nevertheless, it is necessary to note that, in our study, we analysed household’s differences in consumption choices within a “generationalist framework”, that is, we focused on studying the impact on consumption of belonging to Y- or Z- generation. It is still to be investigated the role played by social changes alone, for instance, by comparing consumption patterns of different cohorts of people (teenagers, possibly) who belong to the same generation at different points in time. However, such an analysis could not be run with our data, which only cover the time period 2001–2013.

In conclusion, the present study makes an important contribution to understand the joint influence of children’s generations and social changes on the expenditure patterns of households for leisure activities. The reported analysis indicates that the role of the generation of children in determining consumption patterns on leisure is significant only when the time difference between generational cohorts is long enough that social, economic, political and technological changes occurred in between can affect people’s lifestyle.

6 Limitations and Future Directions

We acknowledge various limitations in our study. Our sincerest wish, however, is that such limitations and especially the open questions will lead to future research in leisure consumption from a generationalist perspective.

First, we are aware of the difficulty to empirically frame a labile concept as “generation”, especially for the new generational cohorts. In fact, for the previous generations (Traditionalist or Silent Generations, Baby Boomers, and X Generation), it is quite evident to identify which historical events occurred to transform a cohort in a generation and which social changes people experienced (Great Depression and World War II, Cold War and social conflicts in 1968, Fall of Berlin Wall and falling birth rate, respectively). On the contrary, it can be more difficult to frame the generations considered in the present study, Y generation (so called Millennials) and Z generation, in a more specific historical period. Such a difficulty is due to the rapid sequence of socio-economic changes and epochal events that involved the global population in the last decades. Nevertheless, there are some evidences that allow to distinguish between Y and Z generations in terms of attitudes, view of the world and, as the present study suggests, consumption choices. One might be tempted to interpret the differences between Y- and Z- generation expenditure’s patterns revealed in the present work as a mere effect of the different cohorts sampled, 2001 and 2012, without invoking a role played by the generation of offspring. That is indeed also a reasonable interpretation of our results. However, we are not aware of any empirical study that compare consumption patterns of people with the same age, but belonging to different generations in two separated time-windows, which could support or dismantle our interpretation of empirical results. Indeed, all of the studies we are aware of in the vast literature on generational attitudes and consumptions (Brosdahl and Carpenter 2011; Chhetri et al. 2014; Kolnhofer-Derecskei et al. 2017) clearly highlight differences between generational cohorts, but focus on consumption occurred in the same year, by also making it difficult to untangle the effects of age and generation. Unfortunately, empirical studies that could help to better untangle the interplay between generations and the inter-time between sampled cohorts could not be done through our data. In fact, such studies would require one to consider the consumption of teenagers’ cohorts (13–17 years old) belonging to the same generation in 2012 and 2023 (Z-generation offspring in both years) or the consumption of teenagers’ cohorts belonging to the same generation in 1996 and 2007 (Y-generation offspring in both years). By involving a time difference between cohorts, which is exactly equal to the one considered in this paper, such hypothetical studies would allow to compare the behaviour of cohorts that belong to the same generation. Therefore, if no difference would emerge in these studies between the cohorts, that would support the relevance of the concept of generation to interpret differences in consumption between households, whereas, if significant differences would arise, that would support the reverse, that is, the prominence of the role played by the inter-time between cohorts with respect to the generation.

Second, the list of expenditure’s variables used to infer families’ style of consumption could be easily enlarged and specified to better highlight the differences between the generations of offspring. On the one hand it’s true that our dataset is informative about a wide range of leisure activities–from sport to painting courses, from musical instruments to events–and allows one to distinguish between goods and services that may or may not be affected by economic changes and technological innovations. On the other hand, however, the database does not include items that surely would enrich the analysis, such as goods and services that appeared in the contemporary era and quickly became iconic of leisure time, such as wellness centres, video games, video on demand services etc. Furthermore, our dataset does not include information about relevant features of goods, e.g., brand and typical cost of a unit.

Third, we have focused our analysis on households in a specific phase of family life-cycle, i.e., families with adolescent offspring. Our choice is motivated by the following considerations. Teenagers are more inclined to be affected by social changes–especially the technological ones–and, as a consequence, to quickly adopt new styles of consumption (Mannheim 1928; tr. En. 1952), which may also influence consumption preferences of other family members. Households with adolescents can be considered as “consolidated families” (Solomon et al. 2006) that, on average, already faced with the primary needs of a new-born family (such as buying the family house and children care) and, therefore, can afford to spend more on leisure time activities and related goods. However, this is also an intrinsic limitation of the present study, since it disregards consumption activity of households at other stages of family life cycle, such as new-born families, families with young adult offspring and families without kids.

In light of the above, the future research directions could be oriented towards a greater understanding of the symbolic value attributed to the consumption of free time, perhaps through a qualitative research methodology that brings out the meaning attributed by adolescents to different leisure time consumption practices. For example, it could be interesting to investigate the motivations that push young people to choose specific activities during their free time. Do they depend on a desire for peers’ approval? Or do they derive from the constant Fear of Missing Out some crucial experiences (Przybylski et al. 2013)?

An analysis performed on more recent data could allow one to better consider the impact of new technologies on the style of consumption: how are consumer trends affected by the flourish of mobile phone applications for everything (monitoring physical activity, listening to music, watching tv, booking flights, reading news, taking pictures and making videos)? That, indeed, might be at the root of what emerged from our study about, for instance, the reduced expenditure on analog photography and newspapers in the 2012. We are led to think that some consumptions might increase much more, whereas others might even disappear. For example, is it conceivable that, on average, the purchase of several goods for leisure time, including several expenditure categories in our data, has been replaced by just the purchase of smartphones in 2018?

Moreover, we believe that a comparison of cohorts from the same generation across a wider time-window, e.g., Z-generation adolescents in 2012 and in 2022, would allow to clarify if the differences in consumption patterns reflect more a cohort or a generation effect.

Finally, we would conclude the present work with an open question that we hope will lead to future debates and empirical studies. Nowadays, the speed with which major technological changes occur and their (almost) global diffusion is increasing. Moreover, such changes, including social media penetration, digital devices, artificial intelligence, might deeply influence the social behaviour of people belonging to more than one generation, accentuated by the rapidity of information spread in contemporary age. Such a rapidity of change could facilitate the transition from a generation to the next one in a short period of time, but, at the same time, could make it difficult to draw a line that separates consecutive generations. As Kolnhofer-Derecskei et al. clearly argued (2017):

This lifetime-long generation transition has become much shorter in the case of the recent generations; the quicker the technological innovations are implemented, the more difficult it is to determine the transition between the generations.

This will probably be the challenge that the sociology of generations will have to face in the near future.


  1. 1.

    Most of the social research work as well as of the popular publicity identifies the Y-generation with those born between 1980 and 2000. In this context it was decided to use the categorisation of Nielsen because it was considered more coherent with the literature concerning the identification of different generational groups as consumer actors. Nielsen ranks the Millennials as those born between 1981 and 1994.

  2. 2.

    According to Morin 1962, libidinization is a continued interest and appetite for new goods that satisfy transitory and nonessential needs.

  3. 3.

    According to ISTAT, a household is defined by people that live together, linked by emotional ties, relationship, marriage, affinity or adoption.

  4. 4.

    For more details on the survey design, the methodological note for years 2001, 2007 and 2012 can be downloaded at: (language: Italian).

  5. 5.

    A few families with both parents unoccupied have been removed from the analysis, as well as the few families that showed no expenditure at all among the 25 considered expenditure categories.

  6. 6.
  7. 7.

    Average Linkage method of hierarchical clustering was also performed, giving a similar partitioning of the families in the data set to the Complete Linkage algorithm.

  8. 8.

    More details on the survey design can be found at



  1. Alwin, D. F., & McCammon, R. J. (2003). Generations, cohorts, and social change. In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of the life course (pp. 23–49). New York: Springer.CrossRefGoogle Scholar
  2. Anderberg, M. R. (1973). Cluster analysis for applications. New York: Academic Press.Google Scholar
  3. Bagnasco, A., Barbagli, M., & Cavalli, A. (1997). Corso di sociologia. Bologna: Il Mulino.Google Scholar
  4. Bauman, Z. (2000). Liquid modernity. Cambridge: Polity.Google Scholar
  5. Becker, G. S., & Tomes, N. (1986). Human capital and the rise and fall of family. Journal of Labor Economics, 4(3), 1–39.CrossRefGoogle Scholar
  6. Bhayani, A. (2015). Social and peer influences in college choice. ICEBMM 2015: International conference on economics and business market management, World Academy of Science, Engineering and Technology, United States, 1–20.Google Scholar
  7. Blackshaw, T. (2010). Leisure. New York: Routledge.CrossRefGoogle Scholar
  8. Bourdieu, P. (1991). Language and symbolic power. Cambridge: Polity.Google Scholar
  9. Brosdahl, D. J. C., & Carpenter, J. M. (2011). Shopping orientations of US males: A generational cohort comparison. Journal of Retailing and Consumer Services, 18, 548–554.CrossRefGoogle Scholar
  10. Chhetri, P., Hossain, M. I., & Broom, A. (2014). Examining the generational differences in consumption patterns in south East Queensland. City, Culture and Society, 5(4), 1–9.CrossRefGoogle Scholar
  11. Choukas-Bradley, S., Giletta, M., Cohen, G. L., & Prinstein, M. J. (2015). Peer influence, peer status, and prosocial behavior: An experimental investigation of peer socialization of adolescents’ intentions to volunteer. Journal of Youth and Adolescence, 44(12), 2197–2210.CrossRefGoogle Scholar
  12. Confcommercio, Imprese per l’Italia (2014). Nota di aggiornamento rapporto consumi. Resource document. Ufficio Studi Confcommercio. Accessed 9 Jan 2019.
  13. Corsten, M. (1999). The time of generations. Time & Society, 8(2), 249–272.CrossRefGoogle Scholar
  14. Costanza, D. P., Badger, J. M., Fraser, R. L., Severt, J. B., & Gade, P. A. (2012). Generational differences in work-related attitudes: A meta-analysis. Journal of Business and Psychology, 27(4), 375–394.CrossRefGoogle Scholar
  15. Cox, D. R. (1958). The regression analysis of binary sequences. Journal of the Royal Statistical Society: Series B: Methodological, 20(2), 215–242.Google Scholar
  16. Edmunds, J., & Turner, B. S. (2002a). Generations, culture and society. Buckingham: Open University Press.Google Scholar
  17. Edmunds, J., & Turner, B. S. (2002b). Generational consciousness, narrative and politics. London: Rowman & Littlefield Publishers, Inc.Google Scholar
  18. Eisenstadt, S. N. (1956). From generation to generation: Age groups and social structure. Glencoe: The Free Press.Google Scholar
  19. Eyerman, R., & Turner, B. S. (1998). Outline of a theory of generations. European Journal of Social Theory, 1(1), 91–106.CrossRefGoogle Scholar
  20. Funches, V., Yarber-Allen, A., & Johnson, K. (2017). Generational and family structural differences in male attitudes and orientations towards shopping. Journal of Retailing and Consumer Services, 37, 101–108.CrossRefGoogle Scholar
  21. Gilleard, C. (2004). Cohorts and generations in the study of social change. Social Theory & Health, 2(1), 106–119.CrossRefGoogle Scholar
  22. Grant, I. C. (2004). Communicating with young people through the eyes of marketing practitioners. Journal of Marketing Management, 20(5–6), 591–606.CrossRefGoogle Scholar
  23. Kahle, L. R. (1996). Social values and consumer behavior: Research from the list of values. In C. Seligman et al. (Eds), The psychology of values: The Ontario symposium (vol. 8, pp. 135–151). Mahwah: Lawrence Erlbaum Associates.Google Scholar
  24. Kolnhofer-Derecskei, A., Reicher, R. Z., & Szeghegyi, A. (2017). Acta Polytechnica Hungarica, 14(8), 107–125.Google Scholar
  25. Laslett, P. (1972). Household and the family in the past time. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
  26. Leventhal, R. C. (1997). Aging consumers and their effects on the marketplace. Journal of Consumer Marketing, 14(4), 276–281.CrossRefGoogle Scholar
  27. Lindstrom, M., & Seybold, P. B. (2004). BrandChild: Remarkable insights into the minds of Today’s global kids and their relationships with brands. London: Kogan Page.Google Scholar
  28. Lo Verde, F. M. (2014). Sociologia dello sport e del tempo libero. Bologna: Il Mulino.Google Scholar
  29. Lo Verde, F. M. (2015). Sport e tempo libero nelle regioni italiane. In Istituto della Enciclopedia Italiana (ed.), L’Italia e le sue regioni. L’età repubblicana. Territori (pp. 581–600). Roma: Treccani.Google Scholar
  30. Mannheim, K. (1928). Das Problem der Generationen, in «Kölner Vierteljahres Hefte für Soziologie», VII, pp. 157–185, 309–330 (tr. en. 1952, Essays on the sociology of knowledge, Paul Kecskemeti (Ed.) London, Routledge & Kegan Paul Ltd).Google Scholar
  31. Matthews, B. W. (1975). Comparison of the predicted and observed secondary structure of T4 phage lysozyme. Biochimica et Biophysica Acta, 405(2), 442–451.CrossRefGoogle Scholar
  32. McDonnell, E. M. (2013). Budgetary units: A Weberian approach to consumption. American Journal of Sociology, 119(2), 307–350.CrossRefGoogle Scholar
  33. Morin, E. (1962). L'esprit du temps. Paris: Grasset.Google Scholar
  34. Mundt, M. P., Mercken, L., & Zakletskaia, L. (2012). Peer selection and influence effects on adolescent alcohol use: A stochastic actor-based model. BMC Pediatrics, 12, 115.CrossRefGoogle Scholar
  35. Noble, T. (2000). Social theory and social change. Basingstoke: Palgrave Macmillan.Google Scholar
  36. Przybylski, A. K., Murayama, K., DeHaan, C. R., & Gladwell, V. (2013). Motivational, emotional, and behavioral correlates of fear of missing out. Computers in Human Behavior, 29(4), 1841–1848.CrossRefGoogle Scholar
  37. Rawlins, E. (2006). Mother knows best? Intergenerational notions of fashion and identity. Children’s Geographies, 4(3), 359–377.CrossRefGoogle Scholar
  38. Rowland, V. T., Nickols, S. Y., & Dodder, R. A. (1986). Parents' time allocation: A comparison of households headed by one and two parents. Home Economics Research Journal, 15(2), 105–114.CrossRefGoogle Scholar
  39. Sassatelli, R. (2010). Fitness culture: Gyms and the commercialisation of discipline and fun. Basingstoke: Palgrave Macmillan.CrossRefGoogle Scholar
  40. Sassatelli, R., Santoro, M., Semi, G. (2008). Quello che i consumi rivelano: spazi, pratiche e confini del ceto medio. In A. Bagnasco (a cura di), Ceto medio. Perché e come occuparsene (pp. 165–210). Bologna: Il Mulino.Google Scholar
  41. Solomon, M., Bamossy, G., Askegaard, S., & Hogg, M. K. (2006). Consumer behaviour: A European perspective. Edinburgh Gate Harlow Essex: Pearson Education Limited.Google Scholar
  42. Vago, S. (2003). Social change. Upper Saddle River: Pearson Prentice Hall.Google Scholar
  43. Waldkirch, A., Ng, S., & Cox, D. (2004). Intergenerational linkages in consumption behavior. The Journal of Human Resources, 39(2), 355–338.CrossRefGoogle Scholar
  44. Weinstein, J. (2010). Social change. Lanham: Rowman and Littlefield.Google Scholar
  45. White, J. (2013). Thinking generations. The British Journal of Sociology, 64(2), 216–247.CrossRefGoogle Scholar
  46. Yao, Y.-Y. (2003). Information-theoretic measures for knowledge discovery and data mining. In J. Karmeshu (Ed.), Entropy measures, maximum entropy principle, and emerging applications (pp. 115–136). Berlin: Springer.CrossRefGoogle Scholar
  47. Zhang, Z., & Stewart A. M. (2016). Estimation of standardized mutual information. UNC Charlotte Technical Report, preprint no. 2016–07.Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Dipartimento di Scienze Economiche, Aziendali e StatisticheUniversità degli Studi di PalermoPalermoItaly

Personalised recommendations