Estimation: bivariate probit

4.6Estimation: bivariate probit

Finally the connection of culture events and sport events is examined with the bivariate probit analysis. First the depending variable is binary such that 1 equals “daily”, “several times per week” or “several times per month” and 0 equals “less often” or “never”. The results are presented in table 4-13.

The classification in table 4-13 shows that only some explanatory variables are significant: the age cohorts 45-54 and 55-64. They are classified significantly less often than the group “less often” + “never”. The error term u₁ in the cultural events participation equation and the error term u₂ in the sport events participation equation are correlated: ρ = 0.463, meaning that there is a latent visitor density factor. The classification of the depending variables in table 4-14 has been formed as follows: y₁ = 1 if the response is “daily”, “several times per week”, “several times per month” or “less often” and y₂ = 0 if “never”. This classification (yes/no) is for the both depending variables: “How often during the past 12 months on your leisure did you go to concerts, theatrical performances, art exhibitions, etc.?” and “How often during the past 12 month on your leisure did you go to sport events (ice hockey, football, athletics, motor racing, etc.?”

The results in the table 4-14 reveal that the effect of gender is clear: women go more often to cultural events while men are more active sport events consumers. All education above the level 5 (5 = upper secondary school, 6 = college, 7 = university of applied sciences, 8 = bachelor’s degree, university, 9 = master’s degree) are statistically significant in the culture participation model. The point estimate of the direct marginal effect is larger the higher the education except for the level 7 (university of applied sciences). The age cohorts 35-44, 45-54 and 55-64 are significantly more active in culture participation and less active in sport events. The vital segregation point is around the age 35. At that age consumers choose cultural events in the expense of sporting events. The place of residence (in comparison with northern Finland which is the constant in the equation) is significant in the cultural participation model while in the sporting events model only other southern Finland and western Finland are different than the other areas. Typically the unemployment rate in northern and eastern Finland is higher than elsewhere and this might be the reason for the higher sporting events consumption. Moreover the higher unemployment in combination with the lower educational level in these regions explains that sporting events are favoured in northern and eastern Finland.