I’m having some technical problems preventing this message from going out via email/course notification, so I apologize if you are seeing this multiple times.
Thank you for using TIA to prepare for MAS-I. If you took the exam on Tuesday, I hope you passed! If you didn’t pass, then you might be eligible for a free extension. Our extension policy can be found here:
http://www.theinfiniteactuary.com/files/TIA_Extension_Policy.pdf
The exam was published yesterday, and I’ve posted my first draft handwritten solutions to the exam in the sample exams tab – please forgive me if anything is too hard to make out, and do feel free to ask for clarification. The video solutions for the exam will be posted there over the next few weeks. They will be part of the free samples for the course for the next six months, so you’ll be able to see them there whether or not you are logged into the TIA website. Please let me know if you see anything that needs to be corrected.
I did not really like this exam. I agree with all of the answer choices on the CAS preliminary answer key, but I felt that the exam focused too much on a few topics and ignored others entirely. The exam writers still seem interested in cherry picking single words and formulas from the source materials, and then on other questions, there is little or no relationship between what you have to compute and anything that appears in the sources. It feels like the breadth and the depth of what you are required to know for the exam continue to expand.
I do not have a sense of how the CAS will set the pass mark. I feel as though this one was a bit harder than last Fall’s exam, although I still ended up with roughly 33 problems that I felt were type 1 (work immediately) sort of problems. I think that because of the newness of several topics, either to the syllabus, or things that have just never been tested before, scores are likely to be lower this sitting. But the CAS ignored a faulty problem last Fall and did not adjust the pass mark, and so I don’t really feel like I have any basis for prognosticating about what they will do.
It may be of more broad concern since a couple of people have asked whether or not the statement “no interaction terms” under the GAM portion of the D.6 summary sheet contradicts the falsehood of the correct choice B for #38, “A limitation of GAMs is that interactions cannot be added to the model.” The answer is that I do not feel that it does not, despite the near wording – we can include interactions, but we don’t do it through adding interaction terms that we have used up to the point of reaching the GAM model in the course. On slide 2 of lesson D.6.1 when introducing the GAMs I explained that we do not include interaction terms for GAM models so that the model will remain additive, and this sets the GAM apart from GLMs and General linear models. I also explain at the bottom of that same slide that we can include interactions by using multivariable functions of X_1 and X_2 – this would be something like putting in f(X_1, X_2) in place of f(X_1) and f(X_2), and not by adding interaction terms to the model.
Thanks, again, and best of luck with your continuing studies.