STA1010: Statistical methods for science

10% Weekly quizzes
5% Pre-liminary exercises
15% Assessments (1: 7.5%, 2: 7.5%)
10% Group project
60% Examination

Overall comments: I found this unit pretty doable after SCI1020 (which gives an intro to most of the content of STA1010 sans probability and non-parametric tests). Personally, I found it similar to SCI1020 - dry and boring but everything is very structured with topics and their corresponding lectures, worksheets, set out from the start. I recommend this unit, if you don’t like numbers, like me or if you’re comfortable with statistics/ further and are looking for a WAM booster.

Weekly quizzes (10%): Multiple choice, 10qs, 1 hour, 1 attempt
These were medium in difficulty but definitely possible to full mark especially given the excessive time limit. Most of the questions were partially copied from other American stat exams and since these quizzes were open book, I utilised my research skills well. Make sure, you read the stems well, and sometimes for some spice, they change the numbers up compared to the qs available online, so be a bit careful with the calculations. Overall, I would say you can easily get the 10% if you’ve understood the lecture content/ have the notes right in front of you.

Weekly worksheet (10%): Short answer mini prep tests, due Monday night the week after its corresponding applied class
These are meant to be preparatory worksheets before attending the applied class but due to the online setup – changed them to be submitted after the classes. I would recommend, attending the weekly applied classes as sometimes you can practically finish the entire worksheet just from the tutor explaining each qs and they usually drop heavy hints about the qs they want you to attempt. But you can definitely get away with not attending them if you’ve read and understood the lecture notes. Overall, the applied classes are a massive tool for early exam prep as although they are sometimes considerably hard to stay awake in, during the second hour, where you are left to your own devices to complete the worksheet, you can ask all your qs to the tutor.

Assessments (20%): Short/long answer, due approx. two weeks after released
These test approx. 4 weeks at once, as there’s usually one stem with several questions following it, corresponding to each week. This is when the lectures come in handy, as sometimes they do the exact same qs from future assignments along with a clear step by step working out, so make sure not to miss out on these lucky eggs. It's relatively easy to do well in these, if you have been consistent with your lectures and worksheets but if you haven’t (I found these pretty difficult without attending the applied classes– so make sure you start attempting the questions, a week before its due), there are plenty of online resources such as the supplementary videos and online basic stats courses, that give you a step by step for common qs.

Project (10%): 3 parts each with its own due date, spread throughout the semester
Essentially, you have to find a real-life example (such as coins/ the no. of chocolates in a pack) and create an experiment with a hypothesis (eg: fantasy books are rated higher than sci-fi novels), where you can apply a specific hypothesis test. This is an easy 10%, even if you have a bad group, just find accurate data and the descriptive stats on Excel, then perform confidence intervals + hypothesis tests (usually t-test) -> reach a conclusion which does/doesn’t (provide errors if it didn’t) support the hypothesis.

Exam (60%): Multiple choice, 100qs, 3 hrs, closed book and invigilated (eek)
This MCQ exam was unnecessarily given 3 hours but I’m not complaining as I believe, usually its short answer. It’s quite easy to be ready for the exam but only if you had done adequate preparation. A key tip is to make sure; you have completed the mocks as they are relatively representative of the difficulty of the real thing. My tips would be to keep up with the weekly content and attend the applied class, watch the lectures live if possible, and if not, making sure to watch the recordings. In terms of content: attend the very last lecture for exam tips, they list the specific topics that are examined in the exam (mostly the different types of tests, probability).
Tip: I suggest creating a one-note page of a timeline of sort and compiling all the assessment and quiz dates and highlighting whenever a major assessment is due.
Overall, this unit is neatly presented and despite sometimes being boring, you can perhaps seek motivation from the fact that you can do really well with little effort.
Goodluck!

Dr Daniel McInnes (unit coordinator)
And a bunch of Tas

Yes, 2 mock exams provided

3/5

All lectures were live-streamed as well as recorded, the applied classes were streamed via zoom and recorded

um, I don’t think there was a textbook – instead, they had a book with all the course material – lectures, lab worksheets, prelim. exercises

3 x 1hr lectures weekly
1 x 2hr applied class weekly

2020, semester 2.

91 HD

10 On-line Quizzes - 1% each, Labs, prelabs & participation - 10%, Two assignments and group inference project - 20%, Exam - 60%

I should start off by saying that Jon was a great lecturer and the rating should not reflect on him. It was more that the content he had to teach wasn't that great. If you did further and Methods during yr 12, then the first 6 weeks will be a piece of cake, you will learn next to nothing. Basically take a watered down version of the worst part of methods, probability and combine it with the driest part of of further, statistics. After that you learn a few more concepts (Central Limit Theorem, Confidence Intervals) before moving onto hypothesis testing, which is the meat of the course. In itself it's not too hard, but just like Yr 12 probability, the hardest part is picking which formula to build and apply to the situation.

Although this is an easy unit, it requires time throughout the semester, it sucked up time that I wanted to use for other units, you have all these small assessments here and there, prelabs, on-line tests, assignments. Although, as a result of being lax for this unit (only 2 hrs study during swotvac), I probably lost a few more marks than I planned to on the exam. But that was because of prioritising other units, but the message here is that yes it is easy, yes you can slack off, but don't completely slack off.

Dr Jonathan Keith

Yes, 2 with solutions.

2 Out of 5

Yes, with screen capture

You really don't need it.

3x1 hr lectures, 1x2 hr "Lab" (using excel/tute questions)

Semester 1 2013

90 - HD

Not a fan of this unit, really. I found it to be extremely dry, and very broad in terms of difficulty (i.e. some areas were quite challenging, whilst others were ridiculously easy). This unit gives some very useful knowledge and techniques relating to scientific statistical practices and inference techniques, but it is very non-mathematical (if you can't do it on a basic scientific calculator, it's not in the course; e.g. no PDFs/CDFs/Calculus).

Jon is a good lecturer, and he delivers the necessary explanations well. However, he is quite softly-spoken so the attention tends to wander during the lectures; I found it worthwhile to relisten to critical lectures online. In all honesty, I stopped going to lectures after the sixth one (out of 31 or so), but I made sure to rewatch any lecture on things I wasn't sure about during SWOTVAC.

Labs were pretty much useless, in that the tutor simply walks you through the problems in that week's entry of the lab manual. In a way, it does reinforce the practices presented in the lecture, but more through repetition than anything else. You do get marked for attendance and participation (10% of the overall mark, if I understand correctly). The lab manual questions can provide a source of practice exam questions if needs be, I guess.

In-semester assessment consists of:

• Weekly Moodle quizzes - I found these to be quite tricky, but that's often because I was behind on theory. In hindsight, very doable if you understand the material (even cursorily - it's very much a 'plug-and-play' unit in some ways)
• Statistical inference group project - you work in groups of 3-5 people to design a statistical inference project based on a data source of your choice. You'll perform the experiment, collect the data and perform statistical inference tests. This project is broken up into 3 sections over the course of the unit: A - Preliminary report (aim, hypotheses, method, apparati etc), B - Testing and data collection and C - inference. This project is quite easy to do well in, even with little effort.
• 2x assignments - these can also be quite tough if you're not clear on theory, but also very doable if you are.

The areas of study covered (generally) are:

• Creating different types of experiments and procedures thereof (SRS, blocking, stratified random sampling)
• Modes of data collection and analysis
• Population parameters (mean, SD and median)
• Linear regression
• Power, exponential and linear relationships in data plots and residual plots
• Differences between population parameters and sample statistics
• Probability - basic Bayesian probability, manipulation of conditional probability, etc
• Binomial, Normal and Poisson distributions
• Standardisation
• Confidence intervals - various types
• Hypothesis testing - various types
• ANOVA and Chi-squared testing
• Approximations to normality

amongst other areas of study.

Be very prepared to be using a lot of statistical tables...

It's not really my thing, but some of the content (specifically hypothesis testing, ANOVA and confidence intervals) were very interesting and seem to be useful in many (if not most) areas of science.

The exam is very simple - the only real revision I did was the 2012 paper, which was extremely similar to the 2013 one. Success! The questions are very straightforward, and the exam is very easy to prepare for.

Also, from what I hear, this unit is compulsory (unless another Maths credit is done) for those undertaking the B.Sc. It must be completed at some point in the degree.

Dr. Jonathan Keith

Yes. This semester we were only given the 2004 and 2012 exams (with solutions).

2.75-3/5

Yes, with screen capture.

Semester 1, 2013

HD