tag:blogger.com,1999:blog-2191533284937327485.post2474242818595630315..comments2023-10-29T08:22:43.432+00:00Comments on That doesn't seem to add up: Why don't we sample more?John Fabenhttp://www.blogger.com/profile/01949065981896055898noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-2191533284937327485.post-25421196289655168542010-05-17T23:23:23.842+01:002010-05-17T23:23:23.842+01:00Haha, interesting. Although doesn't that mean ...Haha, interesting. Although doesn't that mean we should only respond to some 999 calls (sampled at random) too?Andyhttps://www.blogger.com/profile/10447350470836815641noreply@blogger.comtag:blogger.com,1999:blog-2191533284937327485.post-91774272196835484332010-05-17T08:59:26.854+01:002010-05-17T08:59:26.854+01:00I'm willing to have general elections resolved...I'm willing to have general elections resolved using random sampling as long as the government is willing to collect tax the same way.T_Beermonsterhttps://www.blogger.com/profile/02844293558621285358noreply@blogger.comtag:blogger.com,1999:blog-2191533284937327485.post-71259491433063281282010-05-16T01:12:05.611+01:002010-05-16T01:12:05.611+01:00Well, yes, but then we might pick their best quest...Well, yes, but then we might pick their best questions as well. Random sampling increases the variance slightly, but there is already some variance in coursework assessment anyway (we don't ask coursework questions about every possible aspect of every possible course) - is there any reason to believe that the amount of variance we have now is optimal?John Fabenhttps://www.blogger.com/profile/01949065981896055898noreply@blogger.comtag:blogger.com,1999:blog-2191533284937327485.post-28523372394158439622010-05-15T18:55:29.105+01:002010-05-15T18:55:29.105+01:00I guess the problem is that the student might be r...I guess the problem is that the student might be right. Over 10 courseworks, you might actually always mark his worst, right? However, over a large number of courseworks, I'd expect you to mark as many of his "best" as his "worst".Andyhttps://www.blogger.com/profile/10447350470836815641noreply@blogger.comtag:blogger.com,1999:blog-2191533284937327485.post-90601842161748949372010-05-15T12:22:23.481+01:002010-05-15T12:22:23.481+01:00I marked a course once where I was the only marker...I marked a course once where I was the only marker, and the only instruction I was given was to mark about 3 questions (of ~12).<br /><br />The students didn't really like it. I don't understand why, because I was also the tutor, and had they paid attention they could have picked up a pretty good idea of which questions I liked/disliked and saved themselves effort.Broghttps://www.blogger.com/profile/14185464573529387638noreply@blogger.com