| ncaa_rpi | R Documentation |
Wooldridge Source: Data on NCAA men’s basketball teams, collected by Weizhao Sun for a senior seminar project in sports economics at Michigan State University, Spring 2017. He used various sources, including www.espn.com and www.teamrankings.com/ncaa-basketball/rpi-ranking/rpi-rating-by-team. Data loads lazily.
data('ncaa_rpi')
A data.frame with 336 observations on 14 variables:
team: Name
year: Year
conference: Conference
postrpi: Post Rank
prerpi: Preseason Rank
postrpi_1: Post Rank 1 yr ago
postrpi_2: Post Rank 2 yrs ago
recruitrank: Recruits Rank
wins: Number of games won
losses: Number of games lost
winperc: Winning Percentage
tourney: Tournament dummy
coachexper: Coach Experience
power5: PowerFive Dummy
This is a nice example of how multiple regression analysis can be used to determine whether rankings compiled by experts – the so-called pre-season RPI in this case – provide additional information beyond what we can obtain from widely available data bases. A simple and interesting question is whether, once the previous year’s post-season RPI is controlled for, does the pre-season RPI – which is supposed to add information on recruiting and player development – help to predict performance (such as win percentage or making it to the NCAA men’s basketball tournament). For the binary outcome that indicates making it to the NCAA tournament, a probit or logit model can be used for courses that introduce more advanced methods. There are some other interesting variables, such as coaching experience, that can be included, too.
Used in Text: not used
http://www.cengage.com/c/introductory-econometrics-a-modern-approach-7e-wooldridge
str(ncaa_rpi)