Family Structure: Living with Children. Living with a kid should be controlled in this study because of its potential correlation with both an individual’s animation viewing habit and marital status. This correlation of children with both our dependent variable and our variable of interest can produce an omitted variable bias problem if kid is omitted. In other words, without such control, we can wrongly attribute the effect of having a child in a home to animation consumption. Table 1 shows that 19% of the respondents in the study lived with at least one young kid in 2008 whereas about 20% lived with at least one teenager. Kids are huge consumers of animation in Japan and in most parts of the world. In the past, most animation was geared towards children and teenagers. Currently, animation geared more towards adults is becoming common (Xxxxxx, 2005; Price, 2001; Xxxxxxxx, 2000). The 2008 JGSS does not require respondents to specify the motivation for their consumption of animation. If an adult is living with his or her own children, he or she is more likely to be exposed to animation indirectly through their children and is hence more likely to consume more animations (Xxxxxxxx, 2013; Xxxxx, 2006). As noted above, the presence of a kid in a family can also correlate with an individual’s marital status given that most of the children in Japan are born within a marriage. Given this possible correlation, controlling for the presence of a young kid or teenagers in a family is important because without such treatment, the estimate of the relationship between an adult’s animation consumption and marital status could be biased. It is also possible that the presence of younger children in the family can have more influence on an adult’s animation consumption. Compared to a teenager, a young kid usually possess less independent living skills and self-control and thus would receive more care from family members. For an adult taking care of children at home, sitting together with a young child and watching the kid’s favorite animation can be a good way to build an emotional bond with the kid. As the impacts of younger and older children in a family could differ, it is more reasonable to control for each of them separately in a model. To control the presence of kids in a family, I constructed two separate dummy variables. The first dummy variable, named “young kid”, equals 1 if an adult is living with at least one kid younger than 10 years old and equals 0 otherwise. The other dummy variable, named “teenag...
