The allocation of time to sports and cultural activities: an analysis of individual decisions.

Muniz, Cristina ; Rodriguez, Placido ; Suarez, Maria J. 等

Introduction

Sports and culture are rarely linked in academic research. While a burgeoning literature exists in both cultural and sport economics, both subjects have generally been studied in isolation despite their clear similarities: Both are forms of leisure time allocation that may be jointly chosen by individuals. This common feature justifies a joint analysis of both activities.

An interesting question addressed in this article is whether these leisure activities tend to be demanded jointly or separately. On one hand, both activities may be considered as a social occasion and a chance to spend time and socialize with others. On the other hand, sports and cultural activities are time-intensive goods and may compete for individual leisure time.

The aim of this research is to analyze the individual's decision to allocate time to sports and cultural activities using the neoclassical consumption theory. In Spain, the literature on sports participation (we include physical activity) from an economic perspective is scarce and, as far as we know, no articles study both of these leisure activities. Garcia et al. (2009) develop a structural model of the allocation of time to sports and leisure, assuming a Constant Elasticity of Substitution (CES) utility function. The current paper builds upon Garcia et al.'s model by adding another use of leisure time: time devoted to cultural activities. Thus, we specify a model that defines three uses of time besides working time: sports, cultural, and other leisure activities. This model leads to a system of three demand equations for time not allocated to work, which are estimated using the Seemingly Unrelated Regression (SUR) method.

The dataset used in this research is "La Encuesta de Empleo del Tiempo" (Spanish Time Use Survey, 2002-2003) conducted by the Instituto Nacional de Estadistica (INE; National Statistics Office) according to the Eurostat guidelines for harmonized European time use surveys. The distinctive feature of this database is a diary that provides information on the activities of each individual across a given day, provided in 10 min intervals.

Literature Review

Sports Economics

There is a wide body of literature on sports economics, including several textbooks and handbooks that analyze various aspects of interest to economists. (1) Moreover, there is growing economic literature that specifically examines sports participation, among which it is worth mentioning Humphreys and Ruseski (2006, 2007, 2009, 2010); Downward and Riordan (2007); Wicker, Breuer, and Pawlowski (2009); and Downward and Rasciute (2010).

The research on mass participation in sports draws upon two major theoretical approaches: the neoclassical theory and the heterodox economics approach. The central feature of the neoclassical models is to assume that agents are rational in the sense that their objective is to maximize their individual utility subject to certain restrictions. Moreover, consumer preferences are generally considered invariant. In this framework, sports (and cultural participation) can be treated as commodity demands or as demands for leisure time, either extending the standard labor supply model or applying Becker's allocation of time theory (Becker, 1965).

Heterodox approaches, on the other hand, comprise psychological theories of consumption, post-Keynesian theory, and sociological theory. They tend to provide a more descriptive account of how decisions are made instead of relying on a theory of behavioral optimization. Consequently, habits and a hierarchical structure of demand may be relevant. In addition, these approaches view preferences as dependent upon experience with consumption, thus reflecting learning-by-doing and the preferences and behaviors of others. As a result, preferences are endogenous and can be linked to lifestyles. (2)

Most of the published studies on sports participation are only empirical, and few theoretical models analyze sports participation. Among those papers with a theoretical framework, Cawley (2004) extends Becker's time use model (1965) to physical activity and specifies the SLOTH model; according to this model, people allocate their time to the following activities: sleep (S), leisure (including physical activity) (L), occupation (O), transportation (T), and home production (H). Humphreys and Ruseski (2006) combine the main aspects of the SLOTH framework with a recreation demand model (McConnell, 1992) and distinguish between participation and frequency of participation. Their model predicts that the economic factors that affect these two decisions work in opposite directions: Income has a positive effect on participation and a negative effect on time spent in physical activity. (3)

Humphreys and Ruseski (2007) follow the neoclassical approach to examine sports consumption in the U.S. and analyze variation in participation and time spent in physical activity across groups of activities. Thus, changes in government spending on parks and recreation increase participation in some activities but reduce it in others. Humphreys and Ruseski (2009) demonstrate the importance of economic factors, such as income and opportunity cost, to explain physical activity and conclude that it can be considered to be a normal good.

Downward's (2007) empirical research explores the choice to participate in sports in the U.K. His results suggest "more support for the predictions of heterodox theories for sport demand than the neoclassical theories" (Downward, 2007, p. 650) because the income effects decrease when sport is disaggregated. Further, he finds that changes in working time are not significant determinants of participation. In contrast, socioeconomic variables such as sex, education, and household composition appear to be the main factors that affect sports participation.

Downward and Riordan (2007) also argue that such results could be consistent with Becker's time allocation model. These authors find evidence that investment in personal consumption capital and social capital--through social interactions--increases participation in sports. As with cultural goods, previous consumption can increase current consumption because of the skills acquired (Downward & Rasciute, 2010). Moreover, Becker's approach is also supported more recently by Breuer and Wicker (2008) and Wicker et al. (2009) in the case of Germany. (4)

In Spain, the analysis of sports participation and physical activity has traditionally had a sociological emphasis. This point is illustrated by the studies of Garcia Ferrando (2001a, 2001b, 2005, 2006), Fernandez-Ballesteros (1998), Fernandez-Ballesteros et al. (2004, 2007), Lagardera Otero (1992), Martinez del Castillo et al. (2005, 2006), Moscoso (2005), and Puig and Soler (2004). In these studies, the authors examine the degree to which sedentary lifestyles have become dominant among the Spanish population. They also study how physical activity or sports contribute to the improvement of individuals' feelings of health, wellbeing, and quality of life.

Lera-Lopez and Rapun-Garate (2005, 2007) and Garcia et al. (2009) are the only works that analyze sports participation from an economic perspective in Spain. Other authors examine the links between participation and expenditure in a Spanish region, although no explicit theoretical orientation is noted.

Cultural Economics

Before discussing the empirical literature on participation in cultural activities, we will briefly overview theories of participation in the arts. Performing arts consist of theater, opera, ballet and modern dance, symphony concerts, and chamber recitals. Since the work of Baumol and Bowen (1966), a majority of studies deal with live performing arts, which motivates our focus on the demand for these cultural activities. (5)

As McCarthy and Jinnett (2001) point out, there are two different approaches to the analysis of participation decisions. First, under the traditional economic approach, individuals behave as rational consumers who seek to maximize their utility given their preferences and price and income restrictions.

Other authors find that tastes play a more active role in consumer decisions, especially in relation to cultural goods among others. Many cultural economists assert that current arts demand is influenced by past arts exposure. Within theories on taste cultivation, some authors believe that past consumption can positively influence present and future consumption through rational addiction (Stigler & Becker, 1977). Aficionados are thereby distinguished from casual attendees. However, others focus on consumption as a learning process. This approach directly opposes the rational addiction approach, wherein the formation of tastes can be perceived as a deterministic process and, thus, consumption is shortsighted (myopic). In the learning-by-consuming approach (Levy-Garboua & Montmarquette, 1996, 2002), the formation of tastes is a dynamic process. In other words, any new experience can change consumer tastes.

In summary, consumption is affected by factors such as prices and income. In addition, as people attain more knowledge about cultural goods, they are more likely to participate in consumption because they gain greater satisfaction than those who do not have that information.

In recent years, the empirical literature has focused on a variety of cultural activities, such as original art objects (paintings, sculpture, and other artifacts), theater, and music. Seaman's (2005) survey about attendance at arts events concludes that, although various and complex models of these activities have been developed, there is no consensus regarding the determinants of arts demand. In other words, many of the cases that have been theoretically postulated have not been corroborated by empirical evidence. In particular, the hypotheses that cultural goods always behave as luxury goods and that education has a greater effect than income (two variables that are difficult to separate in practice and lead to multicollinearity problems) are still controversial. In addition, assumptions concerning the values of demand elasticities have not been agreed upon in the literature.

Despite this lack of consensus, one important dimension of the empirical study of art consumption--not present in sports economics research--is that it explores the relationship between different types of cultural demand. Pommerehne and Kirchgassner (1987), Kurabayashi and Ito (1992), Levy-Garboua and Montmarquette (1996), and Prieto-Rodriguez and Fernandez-Blanco (2000) study the correlation between different genres through various econometric techniques.

Sport and Cultural Demand

The possibility of substitution between various activities has been addressed in Kesenne (1981, 1983), Kesenne and Butzen (1987), and Downward and Rasciute (2010). There are some studies about time-intensive goods, such as sports and performing arts, revealing that these leisure activities are more expensive than others even if the prices are the same because of the shadow price of time (see Loyland & Ringstad, 2010).

There are few studies about the relationship between arts and sports demands. Montgomery and Robinson (2006, 2008) discuss the tradeoffs between attendance at cultural events and other types of events. They estimate a system of regressions that includes the natural log of the number of times an event is attended as the dependent variable. Their results suggest that sports activities compete with art events for attendance. (6)

In the U.S., Kopczynski and Hager (2003) consider the relationship between various individual characteristics and attitudes toward the performing arts. They conclude that regular performing arts attendees also attend other leisure activities, including sporting events, more frequently than others. Finally, Swanson et al. (2008) explore the relationship of six motivations (i.e., aesthetic, education, escapism, recreation, self-esteem, and social interaction) with arts performance attendance. The fact that consumption is positively associated with these motives supports previous findings in the sport-spectator literature.

In Europe, Loyland and Ringstad (2009) explore sports demand in Norway and conclude that sports and most other leisure goods are substitutes. In Spain, Fernandez-Blanco and Prieto-Rodriguez (2001) also examine the substitutability between sports and culture. They estimate a trivariate probit model to determine the probability of attending live sports, listening to music, and attending the cinema. These authors find positive error covariances between all three activities, which indicate that sports do not compete with the consumption of music or cinema. However, they do find substantial differences in the estimated effects of gender and education on live sports attendance and on music and cinema consumption. Thus, they conclude that males and those with less education have much higher sports attendance than females and more educated individuals.

The research presented here specifies and estimates a model of the allocation of time to sports, culture, and other leisure activities. The paper develops and estimates a neoclassical model based on the work of Humphreys and Ruseski (2006), Downward and Riordan (2007), and Garcia et al. (2009). However, the current paper expands upon the previous literature because these earlier papers focus on the time allocated to sports.

Model of Participation in Leisure Activities

We specified a structural model of time allocation and derived the system of time demand equations to be estimated. The model developed in this paper is an extension of the research previously conducted by Garcia et al. (2009), which estimated a system of two typical Marshallian demand functions of time devoted to sports and other leisure practices. Furthermore, in this study, three possible time allocation possibilities were examined besides working time: sports, cultural activities, and other leisure time.

Following Garcia et al. (2009), we assumed a simplified CES as the utility function. The CES utility function is commonly used to model the static labor supply choice. (7) We chose this function for its flexibility--it does not impose a linear relationship between wages and time spent on various leisure activities--and for ease of estimation. However, it is important to note that this functional form does not allow inferior goods to be included in the model. (8) However, this issue is not a serious problem for our research because sports and cultural activities can be considered normal goods.

We assumed that individual preferences are a function of consumption and time that people devote to sports, culture, and other leisure activities. According to the neoclassical model, the purpose of each individual is to maximize the level of utility subject to a budget constraint.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)

where U denotes utility, c is consumption, [l.sub.0] is leisure time (not devoted to sports or cultural activities), [l.sub.1] is time spent on sports activities, [l.sub.2] is time spent on cultural activities, w is hourly earnings, y is non-labor income, and T is the time endowment (24 hours per day). Finally, [alpha], [beta], [delta], and [gamma] are parameters of the utility function that must meet the following conditions:

[gamma] > -1

[alpha], [beta], [delta] > 0

Additionally, we allowed both observable and unobservable factors to enter the preferences through the parameters a, p, and 8 according to:

[alpha] = [e.sup.(z0[phi]0-[mu]0)] (2)

[beta] = [e.sup.(z1[phi]1-[mu]1)] (3)

[delta] = [e.sup.(z2[phi]2-[mu]2)] (4)

where [Z.sub.0], [Z.sub.1], and [Z.sub.2] are vectors of explanatory variables that may influence an individual's well-being, and [[mu].sub.0], [[mu].sub.1], and [[mu].sub.2] are random variables that capture unobservable factors affecting the individual valuation of leisure activities.

From the first-order conditions for interior solutions, we obtained the following system of relative demands:

log [c/[l.sub.0]] = [1/[1+[gamma]]log w - [[Z.sub.0][[phi].sub.0]/[1+[gamma]][gamma] + [[mu].sub.0]/[1+[gamma]] (5)

log [c/[l.sub.1]] = [1/[1+[gamma]]log w - [[Z.sub.1][[phi].sub.1]/[1+[gamma]][gamma] + [[mu].sub.1]/[1+[gamma]] (6)

log [c/[l.sub.2]] = [1/[1+[gamma]]log w - [[Z.sub.2][[phi].sub.2]/[1+[gamma]][gamma] + [[mu].sub.2]/[1+[gamma]] (7)

Estimating Equations (5)-(7) allowed us to understand the determinants of individuals' time allocation decisions. Regarding the relationship between sports and cultural activities, the opportunity cost was the same in both cases--the amount of income that is not received by substituting work for all leisure activities--and, thus, the complementarity or substitutability between these types of leisure was analyzed through the correlation between the residuals of the last two equations ([[mu].sub.1] and [[mu].sub.2]). We noted that, although this is consistent with our theoretical approach, it also reveals a limitation of the model because prices are not included. Most likely, the costs of participation would differ between different cultural activities and sports. (9)

Data and Estimation Procedure

The data used to estimate the model specified in the previous section were provided by La Encuesta de Empleo del Tiempo (The Time Use Survey) conducted by the INE from 2002 to 2003. (10) This survey gathered information on the daily activities of individuals in a sample of approximately 24,000 households. Data collection instruments included individual diaries and questionnaires pertaining to individuals and households. (11)

The sample population in the survey consisted of all members of a household who resided in private homes. The sample was distributed uniformly throughout the year, and the distribution was wider on weekend days because individual behavior is more variable on the weekends.

Individuals' daily activity was recorded by a diary that all household members over the age of 10 completed on a given day. The daily activity diary collected 24 consecutive hours (from 6:00 a.m. until 6:00 a.m. the next day) of data in 10 min intervals. At each interval, the respondent recorded the main activity he or she was engaged in and any secondary activity that was performed at the same time.

These activities are coded according to a list of harmonized activities from Eurostat. The 177 main activities were classified into 10 groups: personal care, work, education, home and family, volunteer work and meetings, social life and entertainment, sports and outdoor activities, hobbies and games, media, and travel and unspecified time uses.

Therefore, we could determine the four dependent variables on the left-hand side of relative demand Equations (5)-(7).

Time spent on sports activities. This variable was defined as the amount of time (measured hours per day) assigned to both sports and physical activity, such as walking, playing football, going to the gym, dancing, fishing, and swimming; it also includes the time spent attending sporting events. It is, therefore, an aggregate variable that measured both active and passive participation in sport activities. There are three reasons for this aggregation. First, a homogeneous definition of sports and culture is necessary. Second, following Dawson and Downward (2009), we considered that both demands generally manifest "a latent variable of sport consumption" (Dawson & Downward, 2009, p. 23)--the same reasoning is applied to cultural activities. Third, aggregation provided us with a larger sample size.

Time spent on cultural participation. This variable was defined in a similar way as the sports variable. It was the amount of time (measured in hours per day) allocated to artistic activities, such as painting, sculpture, ceramics, graphic arts, pottery, making movies, singing, dancing, playing musical instruments, and writing novels or poetry. We also considered the time spent as a spectator at cultural events. Therefore, time spent on cultural participation was also an aggregate variable that included active and passive participation in cultural activities.

Time spent on other leisure activities. This variable was defined as the amount of time (measured in hours per day) assigned to leisure activities other than those captured by the previous two variables. The leisure variable acted as a time residual that captured time not devoted to work, sports, or cultural activities. It is important to note that variables such as time watching television or reading books were subsumed into the leisure residual and were not considered as part of the previous two groups.

Consumption. This variable represented the daily net family income. It was computed using information about monthly family income from all sources. Because income was reported in ranges, we assigned the interval midpoint to each household.

In this research, we only focused on the behavior of the working-age population. Therefore, we estimated the model using a sample of people between 18 and 65 years of age. In addition, to correctly measure individual earnings, we only took into account the survey information regarding each individual's primary occupation. Individuals with a secondary occupation were dropped from the sample. We also eliminated all individuals who worked more than 112 hrs per week and those observations for which no information was available on the variables under study.

Our main goal was to estimate the relative demand system of Equations (5)-(7). This system of relative demand equations was only applicable to the subsample of people who participated in both cultural and sports activities on the day they were interviewed. (12) Because this subsample was not randomly selected from the population, we had to control for possible selection biases. Thus, our estimation procedure consisted of two stages.

First, participation in sports and cultural activities was estimated using a bivariate probit model. The dependent variables in this bivariate probit model were discrete variables that took the value of 1 when individuals participated in the activities during the selected day, and it was zero otherwise. From the bivariate probit coefficients, we computed a transformation of the predicted individual probabilities to correct the selection biases in the demand equations to be estimated in the second stage (see Baffoe-Bonnie, 2009; Ham, 1982).

Second, the linear equation system (5)-(7) was estimated assuming that the random components [[mu].sub.0], [[mu].sub.1], and [[mu].sub.2] were distributed as a trivariate normal distribution with zero means and constant variances. Additionally, our theoretical model imposed the single constraint that the coefficient on the logarithm of wages must be the same in all three equations. Thus, we estimated the system using the SUR method, which allowed correlation between the random terms and cross-equation restrictions on the parameters. Previous studies have shown that this econometric technique can offer more efficient estimators when the equations are linked via their error terms (Cameron & Trivedi, 2005; Greene, 2008). Finally, this equation system was estimated using the subsample of people who participated in sports and cultural activities; therefore, we included the selection rule estimates computed from the previous stage to correct the sample selection problem. (13)

In the empirical analysis, separate equations were estimated for men and women because previous findings in the literature on the allocation of time suggest that there are important behavioral differences depending on gender. Empirical studies on gender differences in leisure time confirm that there is still a leisure gap favoring men. Women experience lower-quality free time and less total free time than men because men have many more hours of pure leisure without other activities or responsibilities, and men's leisure is less likely to be interrupted than women's (Bittman & Wajcman, 2000; Mattingly & Bianchi, 2003; Mattingly & Sayer, 2006). Furthermore, following Swanson et al. (2008), gender may be associated with different motivations in both sports and cultural participation because males are characterized by self-focused goals, while females pursue harmonious relations and affiliation with others in their leisure time.

Table 1 provides the definition and summary statistics of the variables used in the empirical analysis.

The independent variables included in bivariate probit models to explain the probability of participation in cultural and sports activities are as follows: age, age squared, health, marital status, number of children younger than 12-years-old, number of adults in the home, degree of urbanization, non-labor income, region, educational level, and dummies to control for the term and day of the week when the information is collected.

Regarding the relative demands for leisure time, the explanatory variables included are as follows: age, age squared, marital status, number of children younger than 12-years-old, number of adults in the home, a dummy to control for the weekend, predicted hourly earnings, and two terms to correct sample selection biases.

In the system of relative demand equations, we included predicted hourly earnings because wages are computed as a ratio of labor earnings and working hours, which may lead to endogeneity problems. Additionally, there is no information about the earnings of nonworking individuals. (14)

For identification reasons, we assumed that some variables only influenced the decision to participate but not the intensity. (15) In particular, in the SUR estimates, we dropped health status, term, degree of urbanization, and non-labor income because these variables may affect the probability of participating but not the amount of time devoted to the activity. Moreover, educational level and region were assumed to have an indirect influence on the leisure demands via their effect on individual earnings. (16)

Results

The coefficients estimated in the bivariate probit model that explains the probability of participating in sports and cultural activities are presented in Table 2. Although interpretation of the coefficients in probit regression was not straightforward, the marginal effects had the same signs as the regression coefficients for all individuals in the sample.

It is worth noting that the correlation coefficient was positive and significant for both males and females, which means that individual sports participation and cultural participation were not independent. Instead, there was a relationship of complementarity between the two leisure activities.

In Table 2, the estimated coefficients suggested that males and females display the same behaviors in relation to cultural activities and sports according to age. Age displayed a U-shaped relationship to participation in cultural activities and sports (although this relationship was not significant for females' participation in cultural activities). For males, the minimum probability of participation in cultural activities was reached at the age of 45. This result reinforces a general hypothesis about arts participation: "After some age, an evening out can be presumed to entail additional implicit costs" (Gray, 2003, p. 358). This variable showed a negative effect, which was similar to the findings of most previous studies (Downward, 2007; Gratton & Tice, 1991; Humphreys & Ruseski, 2006). Furthermore, the influence of age on sports participation agreed with the results of Garcia et al. (2009).

With regard to other explanatory variables, the male probability of sports participation increased when a chronic illness was present. This result was surprising but may be reasonable because our dependent variable included both active and passive sports. Thus, the disease may promote passive participation as a spectator of sports.

Being married negatively affected participation in cultural activities. However, for sports activities, this variable was only significant for females, and it had a positive sign. Because sports activities are more frequently practiced by males, married women attend sporting events or participate in certain sports activities with their husbands (Montgomery & Robinson, 2008).

In addition to marital status, household composition also affected participation. Family responsibilities reduced the time available for leisure activity. Like sports, cultural participation was time consuming, and the presence of children under 12 years of age in the household had an adverse effect on both types of participation, especially for females (the variable was not significant for males for cultural activities). In addition, the presence of adults in the household also reduced participation in cultural activities.

Education level showed a positive effect on both types of leisure activities. Moreover, if we compared the marginal effects of the explanatory variables, the coefficients suggested that this effect was higher for cultural activities. As we would expect a priori, there is evidence of the importance of the acquisition of skills, which allows increased appreciation and enjoyment of cultural activities (Baumol & Bowen, 1966; Gray, 2003).

The results also showed that non-labor income had a positive effect on both activities (only significant for females), as is suggested by theory. Regarding the effect of the time of year, we found that sports participation intensified during the spring and summer seasons, and cultural participation decreased during these terms. One possible explanation for this trend is that sports activities are held outdoors, so the weather becomes a significant factor, whereas cultural activities are often carried out in enclosed places. Moreover, the practice of these activities is concentrated on weekends, when individuals have more leisure time and lower opportunity costs of participation.

In terms of the geographic variables, it might be expected that the likelihood of sports and cultural participation would be higher in urban areas, especially in larger cities, where there is greater supply of such activities. The results confirmed in some cases--based on sign, although variables are not always significant--that participation in sports and cultural activities is promoted by the presence of major sports and cultural facilities in big cities.

As mentioned before, we estimated three equations: the ratios of consumption to time allocated to sports, cultural activities, and other leisure activities. However, we only showed the results for the relative demands for culture and sports because these are the two equations of interest according to the goal of our research. In Table 3, we show the results for two of the three equations estimated. (17)

A Breusch-Pagan test was used to examine the correlation of errors across equations. The null hypothesis of diagonal errors was rejected because there was a positive correlation between the residuals of the three demand equations. Therefore, time spent on sports, culture, and other leisure activities were not independent. This finding may be caused by the common character of both cultural activities and sports as social activities. Thus, the individual presumably achieved his or her desire to interact with other individuals through participation in one activity or the other.

In accordance with our objectives, our comments exclusively address the results regarding the relative demands for cultural activities and sports. First, we saw that the influence of some variables on the participation decision was different from their effect on the time allocation decision. Therefore, participation and participation frequency are different decisions (Humphreys & Ruseski, 2006).

To analyze the effect of these variables properly, we noted that the signs were interpreted in the opposite direction in the estimates because c/[l.sub.1] and c/[l.sub.2] were the dependent variables.

Thus, the relative demands for cultural activities and sports declined in relation to hourly earnings because, when hourly earnings grew, the opportunity costs of all types of leisure activity increased. (18)

Additionally, we computed a Wald test with the null hypothesis that the earnings coefficients were the same in all equations, and we accepted this hypothesis. Thus, the results were consistent with our theoretical model, and we concluded that the SUR method was more convenient than ordinary least square estimation because it allowed us to incorporate cross-equation restrictions.

A positive effect of age was identified. The relative demands of both males and females continued to increase with age, although age was only significant for cultural activities for females and only for sports for males. There was no consensus in the literature about the effect of this variable; some previous studies also found positive effects for age, while others found the opposite effect.

The presence of children at home reduced the female probability of participation in cultural and sports activities and the relative demand for culture. Thus, for females, children discouraged both the choice to participate and the frequency of participation in cultural activities. Following Gray (2003), although having children at home implies "childcare expenses and implicit costs in the form of parental concern" (Gray, 2003, p. 358), we expected a priori that mothers would like to share cultural activities with their children. However, the results suggested the opposite effect. In previous studies on the frequency of participation in sports, the presence of children reduced sports participation, especially for females (Gratton & Tice, 1991; Humphreys & Ruseski, 2006). However, in Garcia et al. (2009), this factor positively affected the relative demand for sports of both men and women. This finding may suggest that parents practice such activities to ensure that their children develop a taste for sports.

The presence of adults in the household also reduced the intensity of participation in cultural and sports activities, probably because it increased the costs of family entertainment. Marriage undermined the relative demands of males and females for sports and cultural activities. Finally, it is logical that participation in these cultural activities intensified during the weekend for females.

The significance of the correction terms of selection bias ([lambda]) confirms that the time allocation decision was logically dependent on the decision to participate for both activities.

Conclusions

In this paper, we specified and estimated a structural neoclassical model to analyze individual behavior concerning sports and cultural activities and disaggregated the choices of males and females. The dataset came from the Time Use Survey elaborated by the INE for the period of 2002 to 2003.

To estimate the relative demands for the different types of leisure, we applied the SUR methodology. These demands equations [(5)-(7)] were estimated only on the subsample of individuals who allocated time to both sports and culture during the day selected. As this sample selection may bias the results, we controlled for possible selection biases by including two correction terms in the demand equations. These terms were computed from the bivariate probit estimates calculated in the first stage.

A comparison with previous literature is not straightforward because there is no similar homogeneous definition of cultural and sports activities. However, most of our results concerning participation in cultural activities--especially with respect to the effects of education, age, marital status, and number of children--were similar to those found in previous research. Our results also reinforced the findings in the empirical literature on sports participation: Being younger, educated, and having no children at home increase sports participation.

As both sports and culture are time-intensive, family responsibilities negatively affect both participation and participation frequency, especially for females and for cultural activities. Our results suggested the continued presence of a gender gap in leisure: Women are at a substantial disadvantage because childcare is still distributed unevenly.

Therefore, the government could promote greater job flexibility and better work-family policies to enhance sports and culture activities on weekdays. This opportunity cost is higher when wages increase, and the shadow prices for cultural and sports activities increase with an increase in predicted earnings per hour. Moreover, we suggest that many of these activities require travel time; thus, the government should encourage people to be more involved in the practice of sports and build new infrastructure to reduce the gap between rural and urban areas in our model.

Household composition also affected leisure time enjoyed by the individual in other ways. Previous studies show that married people have less leisure time compared to the non-married. Our results supported these findings, showing a negative effect over sports and culture demands--although marriage increased the probability for wives to participate in sports.

However, practicing these activities is not free due to monetary barriers. These barriers increase along with the family size because sports and culture are social activities, often shared with other household members.

The government should also try to eliminate these barriers by subsidizing and increasing grants to low-income families. In particular, as cultural participation is different from general leisure recreation--because of specific training and because skills are often necessary to appreciate cultural goods--educational barriers are especially important in the case of culture, as shown by the calculated marginal effects.

In summary, we found differences in time allocation for sports and cultural activities between males and females and presented evidence for the existence of a gender gap in free time. This result validates our decision to conduct separate estimates based on sex and allows us to suggest actions to be taken by public authorities.

In addition, the incorporation of a new demand equation that measures time allocated to culture allowed us to compare different individual leisure practices. Previous models focused only on sports and did not allow for such comparisons.

Finally, it is worth noting that there was a positive correlation between sports and cultural activities, both in terms of the probability of participation and the amount of time allocated to these activities. This finding suggests that a relationship of complementarity exists between both activities.

References

Ahn, S. C., & Lee, Y. H. (2003). Life-cycle demand for Major League Baseball. Paper presented at the 2003 Western Economics Association meeting, Denver, CO.

Amemiya, T. (1985). Advanced econometrics. Cambridge, MA: Harvard University Press.

Andreff, W., & Szymanski, S. (Eds.). (2006). Handbook on the economics of sports. Cheltenham, United Kingdom: Edward Elgar Publishing.

Baffoe-Bonnie, J. (2009). Black-White wage differentials in a multiple sample selection bias model. Atlantic Economic Journal, 37, 1-16.

Baumol, W. J., & Bowen, W. G. (1966). Performing arts: The economic dilemma. Cambridge, MA: The MIT Press.

Becker, G. (1965). A theory of the allocation of time. Economic Journal, 75, 493-517.

Becker, G., Grossman, M., & Murphy, K. M. (1994). An empirical analysis of cigarette addiction. National Bureau of Economic Research, Cambridge, MA.

Bittman, M., & Wajcman, J. (2000). The rush hour: The character of leisure time and gender equity. Social Forces, 79, 165-189.

Breuer, C., & Wicker, P. (2008). Demographic and economic factors influencing inclusion in the German sports system: A microanalysis of the years 1985 to 2005. European Journal for Sport and Society, 5, 35-43.

Breuer, C., & Wicker, P. (2009). Decreasing sports activity with increasing age? Findings from a 20-year longitudinal and cohort sequence analysis. Research Quarterly for Exercise and Sport, 80(1), 22-31.

Buraimo, B., Humphreys, B., & Simmons, R. (2010). Participation and engagement in sport: A double hurdle approach for the United Kingdom. Selected works of Babatunde Buraimo, University of Central Lancashire, United Kingdom.

Cameron, A. C., & Trivedi, P. K. (2005). Microeconometrics: Methods and applications. Cambridge, MA: Cambridge University Press.

Cawley, J. (2004). An economic framework for understanding physical activity and eating behaviour. American Journal of Preventive Medicine, 27, 117-125.

Dawson, P., & Downward, P. (2009). Participation, spectatorship, and media coverage in sport. Manuscript in preparation, University of Bath, United Kingdom.

Downward, P. (2007). Exploring the economic choice to participate in sport: Results from the 2002 General Household Survey. The International Review of Applied Economics, 21, 633-653.

Downward, P., & Dawson, A. (2000). The economics of professional team sports. New York: Routledge.

Downward, P., Dawson, A., & Dejonghe, T. (2009). Sports economics: Theory, evidence, and policy. London: Butterworth-Heinemann.

Downward, P., & Rasciute, S. (2010). The relative demands for sports and leisure in England. European Sport Management Quarterly, 10, 189-214.

Downward, P., & Riordan, J. (2007). Social interactions and the demand for sport: An economic analysis. Contemporary Economic Policy, 25, 518-537.

Farrell, L., & Shields, M. A. (2002). Investigating the economic and demographic determinants of sporting participation in England. Journal of the Royal Statistical Society (A), 165, 335-348.

Fernandez-Ballesteros, R. (1998). Quality of life: Concept and assessment. In J. Adair, D. Belanger, & K. Dion (Eds.), Advances in psychological science: Social, personal and cultural aspects (pp. 387-406). Hove, United Kingdom: Psychology Press.

Fernandez-Ballesteros, R., Kruse, A., Zamarron, M. D., and Caprara, M. G. (2007). Quality of life, life satisfaction, and positive aging. In R. Fernandez-Ballestros (Ed.), Geropsychology: European perspective for an aging world (pp. 197-223). Cambridge, MA: Hogrefe & Huber Publishers.

Fernandez-Ballesteros, R., Zamarron, M. D., Rudinger, G., Schroots, J. J. F., Hekkinnen, E., Drusini, A., et al. (2004). Assessing competence: The European Survey on Aging Protocol (ESAP). Gerontology, 50, 330-347.

Fizel, J. (2006). Handbook of sports economics research. Armonk, New York, M. E. Sharpe.

Fort, R. (2011). Sports economics (3rd ed.). Upper Saddle River, NJ, Prentice Hall.

Friedberg, L., & Webb, A. (2006). The chore wars. Selected works of Anthony Webb, University of Maryland, College Park, Maryland.

Garcia, J., Lera-Lopez, F., & Suarez, M. J. (2009). Estimation of a structural model of the determinants of the time spent on physical activity and sport: Evidence for Spain. Economic discussion paper, Universidad de Oviedo, Asturias, Spain.

Garcia Ferrando, M. (2001a). Los Espanolesy el deporte: Practicasy comportamientos en la ultima decada del Siglo XX. [The Spanish and sport: Practices and behaviors in the last decade of the 20th Century.] Madrid, Spain: Ministerio de Education, Cultura, y Deporte, Consejo Superior de Deportes.

Garcia Ferrando, M. (2001b). La gestion del deporte en el ambito municipal: De la promotion a la fidelizacion del cliente. [Sport management in the municipal area: Promotion to the customer loyalty]. In M. Latiesa et al. (Eds.), deportey cambio social en el umbral del Siglo XXI (pp. 17-40). Madrid, Spain: Librerias Deportivas Esteban Sanz.

Garcia Ferrando, M. (2005). Globalization y deporte: Paradojas de la globalization. [Globilization and sport: Paradoxes of globalization]. In A. Arino (Ed.), Las encrucijadas de la diversidad cultural (pp. 453-466). Madrid, Spain: Centro de Investigaciones Sociologicas.

Garcia Ferrando, M. (2006). Posmodernidad y deporte: Entre la individualization y la masificacion [Postmodernism and sport: Between the indiviualization and massification]. Encuesta de los Habitos Deportivos de los Espanoles 2005. Madrid, Spain: CIS/Siglo XXI.

Gratton, C., & Tice, A. (1991). The demand for sport: A two-stage econometric model estimated from the health and lifestyle survey. Discussion paper, Manchester Polytechnic, United Kingdom.

Gray, C. M. (2003). Participation. In R. Towse (Ed.), A handbook of cultural economics (pp. 356-365). Cheltenham, United Kingdom: Edward Elgar.

Greene, W. (2008). Econometric analysis (6th ed.). New York: Prentice Hall.

Ham, J. C. (1982). Estimation of a labour supply model with censoring due to unemployment and underemployment. Review of Economic Studies, 49, 335-354.

Heckman, J. J. (1979). Sample selection bias as a specification error. Econometrica, 47, 153-61.

Hovemann, G., & Wicker, P. (2009). Determinants of sport participation in the European Union. European Journal for Sport and Society, 6, 51-59.

Humphreys, B., & Ruseski, J. (2006). Economic determinants of participation in physical activity and sport. Manuscript in preparation, University of Illinois, Urbana-Champaign.

Humphreys, B., & Ruseski, J. (2007). Participation in physical activity and government spending on parks and recreation. Contemporary Economic Policy, 25, 538-552.

Humphreys, B., & Ruseski, J. (2009). The economics of participation and time spent in physical activity. Manuscript in preparation, University of Alberta, Canada.

Humphreys, B., & Ruseski, J. (2010). The economic choice of participation and time spent in physical activity and sport in Canada. Manuscript in preparation, University of Alberta, Canada.

Humphreys, B., Soebbing, B., & Lee, Y. (2010). Consumer behaviour in lottery: The double hurdle approach and zeros in gambling survey data. International Gambling Studies, 10, 165-172.

Instituto Nacional de Estadistica (2004). La Encuesta de Empleo del Tiempo [Spanish Time Use Survey 2002-2003]. Madrid, Spain: Author.

Johnston, J., & Dinardo, J. (1997). Econometric methods (4th ed.). New York: McGraw-Hill.

Juster, F. T., & Stafford, F. P. (1991). The allocation of time: Empirical findings, behavioral models, and problems of measurement. Journal of Economic Literature, 29, 471-522.

Kesenne, S. (1981). Time-allocation and the linear expenditure system. Recherces Economiques de Louvain, 46, 113-123.

Kesenne, S. (1983). Substitution in consumption: An application to the allocation of time. European Economic Review, 23, 231-239.

Kesenne, S., & Butzen, P. (1987). Subsidizing sports facilities: The shadow price-elasticities of sports. Applied Economics, 19, 101-110.

Kopczynski, M., & Hager, M. (2003). Denver: Performing arts research coalition community report 2002. Denver, CO: Pew Charitable Trusts.

Kurabayashi, Y., & Ito, T. (1992). Socio-economic characteristics of audiences for western classical music in Japan: A statistical analysis. In R. Towse & A. Khakee (Eds.), Cultural economics (pp. 275-287). Heidelberg, Germany: Springer-Verlag.

Lagardera Otero, F. (1992). Deporte y calidad de vida: La sociedad deportivizada [Sport and quality of life: The sportsmanship society]. In C. Garcia Lopez (coord.), Actas del 80 Congreso Cientifico Olimpico 1992 (pp. 412-423). Benalmadena, Spain: Instituto Andaluz del Deporte, Junta de Andalucia.

Lee, Y. H., & Smith, T. G. (2008). Why are Americans addicted to baseball? An empirical analysis of fandom in Korea and the United States. Contemporary Economic Policy, 26, 32-48.

Leeds, M., & Von Allmen, P. (2002).The economics of sports. Boston, MA: Addison-Wesley.

Lera-Lopez, F., & Rapun-Garate, M. (2005). Sports participation versus consumer expenditure on sport: Different determinants and strategies in sports management. European Sports Management Quarterly, 5, 167-186.

Lera-Lopez, F., & Rapun-Garate, M. (2007). The demand for sport: Sport consumption and participation models. Journal of Sport Management, 21, 103-122.

Levy-Garboua, L., & Montmarquette, C. (1996). A microeconometric study of theater demand. Journal of Cultural Economics, 20, 25-50.

Levy-Garboua, L., & Montmarquette, C. (2002). The demand for the arts. Presented in a scientific series at Centre Interuniversitarie de Reserche and Analyze des Organizations, Montreal, Canada.

Loyland, K., & Ringstad, V. (2009). On the price and income sensitivity of the demand for sports: Has Linder's disease become more serious? Journal of Sports Economics, 10, 601-618.

Loyland, K., & Ringstad, V. (2010). Borrowing demand at local public libraries in Sweden. Presented at the 16th International Conference on Cultural Economics by the Association for Childhood Education International, Copenhagen, Denmark.

Maddala, G. S. (1983). Limited dependent and qualitative variables in econometrics. Cambridge, England: Cambridge University Press.

Martinez del Castillo, J., Jimenez-Beatty, J. E., Graupera, J. L., & Rodriguez, M. L. (2006). Condiciones de vida, socializacion, y actividad fisica en la vejez [Living conditions, socialization, and physical activity in old age]. Revista International de Sociologia, 44, 39-62.

Martinez del Castillo, J., Vazquez, B., Graupera, J. L., Jimenez-Beatty, J. E., Alfaro, E., Hernandez, M., & Avelino, D. (2005). Estudio sobre la actividad fisicay deportiva de las mujeres del municipio de Madrid: Habitos, demandas, y barreras [Study about sport and physical activity of women in the municipality of Madrid: Habits, demands, and barriers]. Madrid, Spain: Ayuntamiento de Madrid, Area de Gobierno de Empleo y Servicios a la Ciudadania.

Mattingly, M. J., & Bianchi, S. M. (2003). Gender differences in the quantity and quality of free time: The U.S. experience. Social Forces, 81, 999-1030.

Mattingly, M. J., & Sayer, L. C. (2006). Under pressure: Gender differences in the relationship between free time and feeling rushed. Journal of Marriage and Family, 68, 205-221.

McCarthy, K. F., & Jinnett, K. (2001). A new framework for building participation in the arts. Santa Monica, CA: Rand.

McConnell, K. (1992). On-site time in the demand for recreation. American Journal of Agricultural Economics, 74, 918-25.

Montgomery, S. S., & Robinson, M. D. (2006). Take me out to the opera: Are sports and arts complements? Evidence from the performing arts research coalition data. International Journal of Arts Management, 8, 24-37.

Montgomery, S. S., & Robinson, M. D. (2008). Dick and Jane go to the ballet or was it the baseball game? Empirical evidence on the battle of sexes within a human capital appreciation model. Presented at the 15th International Conference on Cultural Economics by the Association for Cultural Economics International, Northeastern University, Boston, MA.

Montgomery, S. S., & Robinson, M. D. (2010). Empirical evidence of the effects of marriage on male and female attendance at sports and arts. Social Science Quarterly, 91(1), 99-116.

Moscoso, D. (2005). La construction social y cultural del liderazgo en el deporte. [Social construction and cultural leadership in sport]. Apunts: Education Fisica y Deportes, 79, 5-12.

Mulligan, C. B., Schneider, B., & Wolfe, R. (2005). Non-response and population representation in studies on adolescent time use. Electronic International Journal of Time Use Research, 2, 33-53.

Pommerehne, W., & Kirchgassner, G. (1987). The impact of television on the demand for cinema and theatre performances. In N. K. Grant et al. (Eds.), Economic efficiency and the performing arts (pp. 44-61). Akron, OH: Association for Cultural Economics.

Prieto-Rodriguez, J., & V. Fernandez-Blanco (2000, May). Are live sports substitutes of culture consumption? Some evidence for the Spanish Case. Presented at the ACEI 11th International Biennial Conference, Minneapolis, MN.

Prieto-Rodriguez, J., & Fernandez-Blanco, V. (2000). Are popular and classical music listeners the same people? Journal of Cultural Economics, 24, 147-164.

Puig, N., & Soler, S. (2004). Mujer y deporte en Espana: Estado de la cuestion y propuesta interpretativa [Women and Sport in Spain: Current situation and proposed interpretive]. Apunts: Educacion Fisica y Deportes, 76, 71-78.

Sandy, R., Sloane, P., & Rosentraub, M. (2004). The economics of sport: An international perspective. New York: Palgrave Macmillan.

Seaman, B. A. (2005). Attendance and public participation in the performing arts: A review of the empirical literature. Manuscript in preparation, Georgia State University, Atlanta, GA.

Spenner, E. L., Fenn A. J., & Crooker, J. R. (2004). The demand for NFL attendance: A rational addiction model. Manuscript in preparation, Colorado College, Colorado Springs.

Stern, N. (1986). On the specification of labour supply functions. In R. Blundell & I. Walker (Eds.), Unemployment, search, and labour supply (pp. 143-189). Cambridge, United Kingdom: Cambridge University Press.

Stigler, G. J., & Becker, G. S. (1977). De gustibus non est disputandum [Taste of a dispute]. American Economic Review, 67, 76-90.

Stratton, M., Conn, L., Liaw, C., & Connolly, L. (2005). Sport and related physical activity: The social correlates of participation and non-participation by adults. Presented at a Sport Management Association of Australia and New Zealand Conference, Canberra, Australia.

Swanson, S. R., Davis, J. C., & Zhao, Y. (2008). Art for art's sake? An examination of motives for arts performance attendance. Nonprofit and Voluntary Sector Quarterly, 37, 300-323.

Throsby, D. (1994). The production and consumption of the arts: A view of cultural economics. Journal of Economic Literature, 32, 1-29.

Wicker, P., Breuer, C., & Pawlowski, T. (2009). Promoting sport for all to age specific target groups: The impact of sport infrastructure. European Sports Management Quarterly, 9, 103-118.

Zabalza, A., (1983). The CES utility function, non-linear budget constraints, and labour supply: Results on female participation and hours. Economic Journal, 93, 312-330.

Authors' Note

We would like to thank Paul Downward and Jaume Garcia for detailed comments and helpful research assistance on earlier drafts of this paper. We are very grateful to two anonymous referees for useful comments on the previous version of this paper. The usual disclaimer applies.

Cristina Muniz [2], Placido Rodriguez [2], and Maria J. Suarez [1]

[1] Fundacion Observatorio Economico del Deporte

[2] University of Oviedo, Spain

Cristina Muniz is an assistant professor and researcher at the University of Oviedo, Spain. She is also a member of the research team of the Fundacion Observatorio Economico del Deporte. Her main research interests are the behavior of participants in sport as well as the analysis of the demand for performing arts.

Placido Rodriguez is a professor and researcher at the University of Oviedo, Spain. He is Director of the Fundacion Observatorio Economico del Deporte and also President of the International Association of Sports Economists (IASE). He is a co-editor with S. Kesenne and J. Garcia of the book Sports Economics after Fifty Years: Essays in Honour of Simon Rottenberg. He is also former President of Real Sporting de Gijon.

Maria J. Suarez is senior lecturer of economics at the Department of Economics of the University of Oviedo, Spain. Her research interests include labor and sports economics.

Endnotes

(1) Downward and Dawson (2000); Leeds and Von Allmen (2002); Sandy et al. (2004); Fizel (2006); Andreff and Szymanski (2006); Downward, Dawson, and Dejonghe (2009); and Fort (2011).

(2) For the attendance-only case, researchers are beginning to support the notion that fans display characteristics of rational addiction (as in cultural goods, as discussed below) in their consumption behavior. See Becker, Grossman, and Murphy (1994); Ahn and Lee (2003); Spenner et al. (2004); and Lee and Smith (2008).

(3) More recently, Humphreys and Ruseski (2010) also found a similar result: Higher incomes are again associated with higher participation and less time spent on sports, although the income effect is now small.

(4) However, other authors only performed empirical analyses: for example Farrell and Shields (2002), Stratton et al. (2005), Breuer and Wicker (2009), and Hovemann and Wicker (2009).

(5) See Throsby (1994) for an analysis of other art forms.

(6) See also Montgomery and Robinson (2010), who analyze differences in time allocation by sex.

(7) See, for example, Zabalza (1983) and Kesenne (1983). In particular, Kesenne (1983) assumes CES utility functions to analyze substitution in consumption for leisure and non-leisure activities within the Becker framework of the allocation of time. The Stone-Geary function has also been used in consumption studies, but its performance has been disappointing when leisure is incorporated into the set of commodities (Zabalza, 1983). Moreover, in the field of sports economics, Downward and Rasciute (2010) use a Cobb-Douglas utility function, which is more restrictive than the CES because it implies that the elasticity of substitution is equal to 1.

(8) See Stern (1986) for details on the specification of different functional forms for the utility function.

(9) The model developed by Humphreys and Ruseski (2006) contains a full budget constraint that includes time and monetary cost of participation in sport.

(10) The fieldwork was conducted for a full year from October 7, 2002, until October 5, 2003; information was obtained weekly.

(11) Researchers seem to agree that the diary is the best option because the measure of time allocation is more accurate than in other methods. However, diaries also have the disadvantages of sometimes being so general as to be misleading when examining multi-tasking (Juster & Stafford, 1991) and containing biases due to lack of response (Friedberg & Webb, 2006) or inaccurate estimates of time use during a week or a month (Mulligan et al., 2005).

(12) The time spent on other leisure activities also must be positive, but all individuals met this condition.

(13) We can observe a zero value of time allocated to sports or culture either because the person did not participate in that activity the day the information was collected (infrequency of consumption) or because the individual never participated in it. The sample selection model does not take into account these possibilities (Buraimo, Humphreys, & Simmons, 2010; Humphreys et al., 2010). However, the double hurdle model distinguishes both mechanisms that may generate zeros: participation in consumption and corner solutions.

(14) Predicted earnings were computed from a wage equation estimated using the subsample of workers and applying Heckman's two-stage method (Heckman, 1979). The estimation process was not detailed in this paper but is available by request.

(15) According to Maddala (1983), Amemiya (1985), and Johnston and DiNardo (1997), at least one of the explanatory variables in the first equation must be excluded in the second step for identification. However, as Downward and Riordan (2007) point out, finding variables that are excluded from the frequency equation, but not the probability of participation equation, is difficult and arbitrary in many cases.

(16) Downward and Riordan (2007) identify the equations by allowing alternative functional forms in the error terms.

(17) The results of the relative demand for other leisure activities are available upon request.

(18) Downward and Riordan (2007) find the same effect.

Table 1. Summary Statistics of Explanatory Variables
(N = 360)

Variable Definition

Socioeconomic and demographic characteristics

AGE Age of respondent
MARRIED 1 if respondent is married; 0 otherwise
NCHILD12 1 if household includes children aged 12
 years or younger; 0 otherwise
NLABINC Non-labor individual income: Calculated as
 income from other household members.
EDUC1 1 if respondent completed primary education;
 0 otherwise.
EDUC2 1 if respondent completed secondary
 education; 0 otherwise.
EDUC3 1 if respondent completed university
 degree; 0 otherwise.
WEEKEND 1 if day is Saturday or Sunday; 0 otherwise.
ILL 1 if respondent is ill, unfit, or has a
 disability; 0 otherwise.
QUART1 1 if month is January, February, or March;
 0 otherwise
QUART2 1 if month is April, May, or June;
 0 otherwise
QUART3 1 if month is July, August, or September;
 0 otherwise
URB2 1 if respondent lives in a township with
 more than 100,000 inhabitants; 0 otherwise
URB3 1 if respondent lives in a township with
 less than 100,000 inhabitants; 0 otherwise
ADULT3 1 if respondent lives in a household with
 more than 2 adults; 0 otherwise
REGION Dummies for Andalucia, Aragon, Asturias,
 Baleares, Canarias, Cantabria, Castilla-Leon,
 Castilla-La Mancha, Cataluna, Comunidad
 Valenciana, Extremadura, Galicia, Murcia,
 Navarra, Pais Vasco, and La Rioja
LOG W Logarithm of hourly predicted earnings,
 computed from a wage equation through
 Heckman's two-stage method

Variable Mean Standard
 deviation

Socioeconomic and demographic characteristics

AGE 38.8195 14.3404
MARRIED 0.4806 0.5004
NCHILD12 0.2444 0.5936
NLABINC 1242.778 1055.208
EDUC1 0.26944 0.4443
EDUC2 0.3611 0.4810
EDUC3 0.275 0.4472
WEEKEND 0.6228 0.4900
ILL 0.1639 0.3707
QUART1 0.2806 0.4499
QUART2 0.2362 0.4253
QUART3 0.2362 0.4253
URB2 0.0916 0.2890
URB3 0.4195 0.4942
ADULT3 0.4556 0.4988
REGION
LOG W 1.7013 0.3225

Table 2 Bivariate Probit: Cultural and Sports
Participation of Males and Females.

Explanatory Males
variables

 Cultural activities Sports

 Coefficient t Coefficient t

Age -0.0382 -3.59 -0.0525 -7.75
Agesq 0.0421 3.31 0.0750 9.41
Married -0.1626 -2.94 -0.0518 -1.49
Nchild12 -0.0486 -1.38 -0.0565 -2.70
Educ1 0.1752 2.84 0.0428 1.30
Educ2 0.4439 7.08 0.0891 2.47
Educ3 0.5892 8.53 0.2244 5.30
Ill 0.0572 1.07 0.1875 5.86
Weekend 0.4363 11.67 0.4858 19.88
Quart1 -0.1050 -2.03 -0.0329 -1.00
Quart2 -0.0589 -1.13 0.0528 1.59
Quart3 -0.0527 -0.98 0.0861 2.52
Urb2 -0.0165 -0.22 0.0315 0.67
Urb3 -0.0419 -1.01 -0.0785 -2.99
Adult3 -0.0920 -2.18 -0.0308 -1.16
Andalu -0.0780 -1.07 0.0808 1.77
Aragon 0.0561 0.47 0.0263 0.33
Asturi -0.0218 -0.18 0.2131 2.84
Balear -0.0546 -0.36 -0.0602 -0.66
Canari 0.1131 1.07 0.1359 1.81
Cantab -0.1487 -1.20 0.1351 1.81
Castle 0.0408 0.41 0.2825 4.40
Castma 0.0209 0.18 0.0410 0.57
Catalu 0.0719 0.96 0.0076 0.16
Valenc 0.1937 2.12 0.0714 1.16
Extrem 0.1049 0.82 0.3171 3.83
Galici 0.0162 0.19 0.1452 2.66
Murcia -0.2330 -1.70 -0.0461 -0.60
Navarra 0.0387 0.37 0.2455 3.65
Pvasco 0.0331 0.25 0.3557 4.20
Rioja -0.0499 -0.36 0.2178 2.55
Nlabinc 0.0001 0.67 0.0001 1.26

Explanatory Females
variables

 Cultural activities Sports

 Coefficient t Coefficient t

Age -0.0027 -0.27 -0.2798 -4.57
Agesq 0.0011 0.09 0.0408 5.62
Married -0.2028 -4.41 0.0943 3.25
Nchild12 -0.1196 -3.78 -0.0363 -2.00
Educ1 0.1891 3.55 0.0784 2.67
Educ2 0.3261 5.57 0.0141 0.41
Educ3 0,4507 7.19 0.1230 3.15
Ill 0.0296 0.60 -0.0093 -0.31
Weekend 0.4473 13.06 0.2292 10.27
Quart1 -0.0929 -2.00 0.0228 0.76
Quart2 -0.1199 -2.53 0.1127 3.77
Quart3 -0.1289 -2.60 0.1528 4.94
Urb2 -0.0498 -0.73 0.0459 1.09
Urb3 -0.0840 -2.23 -0.0175 -0.73
Adult3 -0.1011 -2.65 -0.0241 -1.01
Andalu -0.0205 -0.31 -0.0113 -0.27
Aragon -0.0354 -0.30 0.1126 1.56
Asturi 0.1562 1.50 0.0594 0.87
Balear 0.1043 0.80 0.0054 0.06
Canari 0.0065 0.06 -0.0246 -0.40
Cantab -0.0065 -0.06 0.2621 3.75
Castle -0.0216 -0.22 0.4017 6.81
Castma 0.0073 0.07 0.0724 1.11
Catalu 0.1177 1.68 0.0089 0.20
Valenc 0.3525 4.35 0.0127 0.23
Extrem 0.2184 1.97 0.3310 4.60
Galici 0.0129 0.16 0.0919 1.87
Murcia -0.1904 -1.42 0.1693 2.40
Navarra 0.1961 2.04 0.3317 5.21
Pvasco 0.1886 1.62 0.4222 5.49
Rioja -0.1068 -0.74 0.2244 2.74
Nlabinc 0.0001 1.26 0.0001 4.61

Table 3. Relative Demands for Sports and Cultural Activities
of Males and Females

Explanatory Males
variables

 Cultural activities Sports

 log c/[l.sub.1] log c/[l.sub.2]

 Coefficient t Coefficient t

Log W 0.9826 6.23 0.9826 6.23
Age -0.0554 -1.85 -0.0175 -0.64
Agesq 0.0412 1.14 0.0041 0.12
Married 0.4852 3.02 0.3646 2.53
Nchild12 0.0732 0.71 0.1137 1.22
Adult3 0.4625 4.35 0.2270 2.38
Weekend -0.1213 -0.81 -0.1781 -1.32
Lambda2 0.0369 0.17 0.1566 0.79
Lambda3 3.2971 2.06 3.2501 2.22
[R.sup.2] 0.1662 0.1778
Sample size 360

Explanatory Females
variables

 Cultural activities Sports

 log c/[l.sub.1] log c/[l.sub.2]

 Coefficient t Coefficient t

Log W 0.8819 7.85 0.8819 7.85
Age -0.0230 -0.85 -0.0521 -1.98
Agesq 0.0125 0.38 0.0386 1.21
Married 0.3372 2.44 0.4754 3.54
Nchild12 0.0757 0.84 0.1563 1.78
Adult3 0.6025 6.01 0.6117 6.27
Weekend 0.0265 0.23 -0.1922 -1.75
Lambda2 0.2039 1.25 0.4879 3.07
Lambda3 1.5682 1.79 2.3614 2.77
[R.sup.2] 0.1924 0.2334
Sample size 389