КУРСОВАЯ РАБОТА TERM PAPER
Title: Peers Effects in Physical Training
Название: Влияние товарищей на занятие спортом
Студент/ Student: Tsay Vladimir
Научный руководитель/ Supervisor:
Yakovlev Evgeny Denisova Irina
In many cases peer effects play significant role but they are not trivial to measure. That is why this issue became popular recently. Most works consider negative peer effects such as excessive alcohol consumption and risky behavior and describe measures to reduce consequences of such undesirable behavior. Instead of direct intervention one may think of indirect influence, for example, substitution ...view middle of the document...
Kremer and Levy used random assignment of peers to track peer effects. Yakovlev exploited complex model with peers’ utilities to get rid of side effects and get pure peer effect. In this paper I follow the letter approach. Peer effects are also widely represented in other fields such as academical studies (see Hoxby), workplace (see Moretti and Mas), in situations where a decision of a person is closely connected with decisions of peers. Participation in sport events is one of these situations. Liu, Patacchini, and Zenou 2011 found significant peer effect for sport activities. In their model, peers influence on individual’s utility. The higher the number of friends is involved in sport, the stronger this influence is. The similar model is shown in Salvy et al. 2008. It considers social involvedness of an individual as a determining factor: a person chooses different levels of physical activity in case of physically active friends and friends with sedentary lifestyle. In this paper I examine key factors for physical activity in Russian population and study the role of peer effects in this context. The following parts contain data description, a model, results and conclusion.
2. Data In my research I use Russian Longitudinal Monitoring Survey (RLMS). This survey is based on various questions about person’s household, education, job, income, health, and other characteristics. Among them there are questions about sport occupations, which are particularly useful for this study. To provide representative sample of Russian population designers of RLMS randomly pick region, then pick district in this region and, finally, a household in this district (to learn more about the methodology please visit RLMS official site http://www.cpc.unc.edu/projects/rlms-hse). Another advantage of this database is large number of observations. Every year interviewers poll more than ten thousand people. This research is based on RLMS Round II database, which includes observations since 1994 up to 2007, excluding 1997 and 1999 years. Overall it yields 120457 observations. Peer definition. How to determine peer connection in such a large sample? The structure of RLMS helps us answer this question. Every person is subscribed to voting district: a small territory unit which contains about 250 people. The key assumption is that people of a same age and gender know each other inside this district. This is our peer group. An average peer group consist of 8 people. To check this assumption I follow Yakovlev 2012. He noticed that people increase their alcohol consumption in the month when they have a birthday, and this increase is statistically significant. In a birthday month alcohol consumption increases on 15% (regression 1 in appendix). If peers within our group were socially connected, their alcohol consumption would increase in case of peer’s birthday too. This actually takes place (regression 2 in appendix). A birthday of one peer (out of 8) increases monthly vodka...