Skicka posten per e-post: Unbiased estimation in seamless phase II/III trials with unequal treatment effect variances and hypothesis‐driven selection rules