Pages

Friday, 17 July 2015

Stats 108: A Review

Introduction -- We are 95% Confident That...

Statistics is, perhaps, not an intuitive course for prospective business students to take. On the other hand, when you think about it, even for a moment, it makes, in some ways, more sense than including the likes of accounting or economics. After all, aren’t those very different disciplines? The one technical, the other scientific. Yet, we’d expect a business graduate to be able to tell you something of what the economy is up to (as in make some meaning of it) or know what a balance sheet looks like. And so it is with statistics. As a discipline statistics can say so much about so many things. In my view, for instance, marketing, as a major, needs a mate to be employable. To economists, understanding the evidence is, one hopes, very important. Thus, one finds statistics.

Note, this course is also known as Stats 101, Stats 101G or Stats 10x but we're looking at it as it appears as a BCom core course.

Aim

To provide an overview of important statistical ideas. Basically that means hypothesis testing, probability, errors and causality in studies, statistical significance (to the extent that this is distinct to hypothesis tests), plots and their interpretations, and a little bit on regression. These ideas mean that the course is necessary for those wanting to do BComs and some majors in other degrees (e.g. psychology).

Model: What does Stats 10x Look Like?

Well, it's three one hour lectures a week with a number of optional tutorials. I never attended a tutorial as I never felt that I needed to but I've heard nothing but good things about them. As Stats 10x is required for all BCom students and is otherwise also important it is an absolutely massive course. There will be many streams and after the test, which finished pretty late, semester one 2015 (I had a test for Stats 20x at the same time) uni was jumping, which should give you an idea of the scale of the thing. Expect the first few lectures to be pretty jam packed but, as ever, things will ease up as some people stop attending. Lectures are recorded. The pacing with the lecturer I had was pretty good. That means Ross, but don't talk: he's funny but he will notice (even if you don't think you're being loud). In some lectures there will be questions to get you thinking, these are answered by clicker which are obtained from the IT place in Kate Edger and can be returned there if malfunctioning (they are loaned like library books just for 12 week periods).

The coursebook is pretty much everything in this course, though. It has all the lecture material with space for notes and even comes with, in the blue pages, revision materials. Those materials can look like further notes (and/or summarised ones, not everything in lecture is in the blue pages) as well as questions that you can try. They are all MCQs as well because the test and exam are entirely multi-choice. I suggest you do look at them because I actually got a question wrong in the test because I didn't know how to do it but I would have had I looked at those questions. It's got a pretty logical structure to it as well but some chapters are substantially shorter than others, although that doesn't mean they're not important.

Assessment

There are two very different types of assessment in this course. They are also delivered in a system with a type of plussage. In Stats 10x the three assignments will be worth (assuming S2 2014's patterns hold in the future) 20%. This is true whatever happens, with the first two (which are pretty easy and quick to do) being worth 5% and the last one 10% (assignment three is much longer and requires SPSS which is only available at uni, I stayed behind late one night... two other dudes had the same idea so we discussed things a little, benefits of a huge course I guess). The test could be worth 10% and the exam 70%, or the test could be worth 20% and the exam 60%. Your final mark will be the higher of two methods. Either the one obtained by 20% + 10% + 70% or the mark from the 20% + 20% + 60% method. I believe that it may be the case that 90% won't necessarily obtain an A+ (although it did for me) because the course is pretty easy to do well in.

However, the assignments and test/exam look very different to each other. Assignments involve some manipulation, judgement, answer generation and interpretation. The test and exam, though, are based on appendices except when they chuck a theory question at you. Thus, they are either theory or interpretation in nature purely (maybe the odd little calculation based on the appendices) within the context of five multiple choice options. Thus, the assessment is pretty chill in stats 10x. Although, just because the two modes of assessment are very different the sorts of thins they include aren't so assignments can be good preparation for the test condition assessments. I guess the most important thing to recall is that all assessments matter and 10% is significant if you bombed on the test, just less so than 20%. Moral? Don't slack because it feels easy, if you screw up, you've still screwed up.

Oh, as a final feature, the assignments are also all marked out of ten. That is, the assignments themselves could be out of, say, 60 but that mark will be converted to a mark out of ten. For instance, 50/60 = 8/10, not 8.3/10. I'm not 100% that I did the rounding the same way the course does but that's the principle.

Content

Well, to be honest, I am not sure I need to explain further. The aim basically encapsulates the important themes: everything else just adorns that. If I did explain further I'd basically be mentioning things in terms of these definitions. These reviews are study materials, they're reviews. Basically, more all additional information in terms of the content would be spoilers.

I will say one thing though. Stats 10x will draw on examples from a wide variety of fields. The thing with statistics courses like Stats 10x is that they are fairly general and, thus, are able to allow one to have snapshots of aspects of disciplines as varied as marketing and ecology. This will appeal to some but not necessarily others. In some cases, the example will be more interesting than the context it comes up in which, usually, is how to interpret something and the validity of those interpretations.

Success: Confidence?

  • Pay attention in the lectures and follow the advice in lectures on how to do the calculations you will have to do in assignments. This will save time, especially with assignment three.
  • Similarly, if the lecture says something is important, it really is. Knowing the boundaries for statistical significance and how to express that matters.
  • When you do your revision, as I said before, read white and blue pages to get all the important stuff from both of them. If practice questions are your thing, do them otherwise you're after the way questions are structured.
  • Multi-choice is impossible to make difficult. Keep calm and remember that. Stats 10x is one of those courses where recognising how the concept is applied and remembering what it is, is pretty much everything.
  • If you are confused about anything, firstly try and think it through yourself. Secondly, ask someone (either in lectures or go to a tutorial) because they are genuinely helpful.

Conclusion?

As you may have gathered, I managed an A+ by the skin of my teeth. I can't recall what I got in the exam and nor do I feel like digging it up but it was really close. It was my only A+ in 2014 (and I tend to be more around the A- level) which is one reason why I feel that Stats 10x is a pretty easy course. I also liked it. It was, perhaps, not quite as interesting as Maths 250 but it was substantially easier (I got a C in Maths 250) and more than interesting enough for me to decide to pursue stats as another major. That gives you both an inkling of my biases in an extremely reflective review (I've since taken and achieved an A- in the also easy Stats 20x) but should also shed some light on the potential, for people sufficiently similar to me, to take to this course. However, the course doesn't really leave one with an overwhelming impression despite being very clear about the stuff it talks about (take note, Infosys 110). That all being said, if I had one thing to say about Stats 10x it is this:

Having done Stats 10x you should be much better prepared to be able to look at a statistic more critically and think about what it really means, although maybe most of that happens at the start of the course.

No comments:

Post a Comment