Hi, There is another thread about this but my query is slightly different: Given the ambiguity of the wording of this question, could the 'survival' referred to here be interpreted as surviving within the course [would that interpretation have been granted full marks]? i.e. both dropping out and passing would result in the student no longer attending the course, so both count as a 'death' in the sense of the estimate. The other thread asks if one or the other could be an interpretation [this is answered by the examiners report; both were given full credit] but my question is on the interpretation of 'Survival'. It's an actuarial exam so interpreting one of the 'Reasons for Leaving' as 'censoring' and the other as 'death' is the prudent thing to do but with the time constraints there's not much time to revise the initial interpretation especially given the intricate ways in which examiners word their questions; this could very easily have been the the first exception where the examiner wanted students to consider both as 'deaths' and neither as 'censored'. Sorry for the long Post. Best, MoleMan
Hi MoleMan, When I first read your post, my initial reaction was that you were just being pedantic - play the game, one item of data is censored, the other is death, come on, stop being silly. Then I read the question again and I think you've made a very good point. It all comes down to what is meant by survival and what information we are seeking... If the language school is interested in how long they take to get students to pass an exam, the death data is passing, censored is dropping out. If the language school is interested in how long before students give up, the death data is dropping out, censored is passing. But what if the language school is just interested in how long before they're not in the class, eg student numbers for planning teachers or classes or whatever? I think this is a valid interpretation of the question. Your problem, as you've already pointed out, is that you've then made the question much easier as it then has no censored data. You would start to worry on exam day that you'll not hit the marking schedule with this - unfair perhaps but let's think about our exam technique... If we do get the "first exception" with 2 different reasons for leaving, no explanation of what "survival" means in the question, and the Examiners report shows both reasons as death data with no censored data in the question, rest assured there will be at least one person (probably from ActEd) making sure that the alternative answer gets full credit. Good luck! John
Thanks! Agreed; my post does sound very pedantic, apologies, it's just that's how I would have interpreted this question at first glance (I panic a bit in exams); examiner is trying something new/different and school is just interested in attendance (as you say; they may need to allocate more resources for more students so it's useful info). It's reassuring; provided I explain the reasoning, that it wouldn't cost too many marks to have interpreted the Q this way.
