Hi! Long term lurker here. I'm super grateful to have gotten into Harvard, Columbia, and Duke for Statistics PhD. I'm interested in Bayesian statistics generally but don't have any specific interests beyond that. I know Columbia has Gelman and Blei but I've heard they turn quite a lot of people down because they're in such high demand. Harvard has a lot of people who do MCMC and theoretical research but I think I'm more interested in Methodological/Applied work which Duke has a high concentration in (Dunson, West, Reiter, etc..). Also Harvard's academic placements seem to be mucchh better than Duke. Does anyone have any advice? Is the obvious choice here Harvard, esp because Duke lost Wolpert + Berger (West also is pretty old? )?