- Points: 30
- Code: M249
- Level: 2

- On this page

This module covers a wide variety of statistical concepts, and staying close to its title, does so from an applied point of view.

The material is outstanding, and it is impressive how difficult concept we introduced without much mathematical background. For example, introducing principal components analysis (PCA) without any matrix algebra was actually clearer than the more theoretical way previous course introduced the concept (at a different university).

The main "issue" with this course is not the developed material, which again is great. The problem with this course is that it does not provide much theory (or any) that allows students to tie together concepts or provide first principle bases. From experience, such background is essential to apply these concepts in a creative and flexible manner and move away from "cookbook" analyses. For example, while the PCA section is fabulous, it really is only an introduction and an additional chapter on the general matrix algebra underlying this would very much strengthen a student's understanding, and increase their ability to use it in new (and potentially different) settings.

I realise this goes beyond the "partial" side of things. But the course is a (mandatory) part of my maths degree. After following multiple courses involving matrix algebra and other concepts that directly underpin these statistical methods it just is very disappointing that the course does not follow through with the actual, more generalizable theory. Closely related, this has been the least arduous module I have followed with the OU (including any of my first year modules), and I had to pace myself not to finish the work in 4 weeks (working only on weekends).

Finally, the software (SPSS) is hardly relevant for modern statisticians or anybody doing statistical analysis. In my day job I supervise medical PhD student who use programming languages such as R or python to analyse their data. I hate to say it but if medical student (who are not trained to program and only get a small degree of statistics training) can use these programs, why does the OU think we can only use a point and click system. In fact because programming (in R for example) is reasonably expressive, computer assignment can be much closer to pen and paper questions. For example, calculating a t-statistic in R:

xbar <- 2

se <- 1.2

tstat <- (xbar-0)/se # assuming the null-hypothesis value for xbar is zero.

Which is identical to variable substitution all student enrolled in this course will know.

So in summary really great material, but it lacks depth, the software is more archaic than pragmatic, and the course is very light (there should be more material!).

Course starting: **October 2019**

Review posted: **August 2020**

This is an OU level 2 introduction module with four practical topics in statistics. At this level many students would struggle with theoretical matrix formulation of PCA as suggested by the comment. SPSS is a decent statistical software package that does a large array of statistical analyses and serves its purpose in M249 by allowing a fairly straightforward application of some of the methods covered without burdening the student with programming. Many non-statisticians using R are more preoccupied with using a method than actually understanding what they are doing and whether that method is appropriate for what they would like to do. Regarding workload, whilst we would like to encourage lots of students to take this module, but nevertheless should warn that few students can complete it in four weeks, working only weekends.

This module is broken into four blocks, entitled Medical Statistics, Time Series, Multivariate Analysis and Bayesian Analysis, each standing alone from the others but built on the statistical concepts developed in M248.

The focus is on building intuition and understanding in the concepts, use and application of statistical techniques across these four broad areas and the material is extremely well presented. The texts do not dig deep into the technical or mathematical details of the methods and techniques but cover them just to the extent needed to convey their workings. This I would think makes the module accessible to a much greater student audience and, generally, I think the materials succeeded in achieving such an aim but never left me wanting for more in order to help me grasp the ideas.

Much of the learning centres around extensive use of a computer running SPSS (and others), which is a great piece of software (and I believe an industry standard). The computing assignments are well structured but to me felt somewhat tedious as one generally had to complete every exercise and example in order to cover every point that was TMA examined, making the module more time consuming than I needed it to be. Fewer, chunkier exercises would have been my personal preference but I do not want to criticise the design of the course on this point as it is, in general, excellent.

There is a lot of material covered in this course and I say this despite my high level of confidence and experience in statistical research (which probably covered around half of the material) but I was often surprised by how many new things I learnt in subject areas I felt I had previously known. This kept me highly engaged with the material and the module. From my perspective I would not recommend sitting M248 and M249 in the same academic year to other students as I do think both modules would be less enjoyable as a result and very demanding on one's time.

The exam was fair but time pressured, assessing across every part of the material without trying to spring any strange surprises.

I would rate the module an 'A'. M249 is made of the stuff that makes it so enjoyable to study and learn with the OU.

Course starting: **October 2014**

Review posted: **October 2015**

The course is varied and falls clearly into separate sections. There is little overlap between medical statistics and time series analysis for example. I found the course both interesting and generally useful and I have two comments:

1. The course is clearly oriented to the practical use of statistical techniques and not to their derivation or theoretical justification. However it sometimes goes too far in just giving you a recipe without even a general explanation of why it works or is justified. In the section on Bayesian statistics, for example, it announces that a computer program is "sampling from a distribution" in such an offhand way I assumed it must be obvious how that was done. In fact it is not obvious at all and I had to go to Google to find out what the computer program was actually doing in general terms. A paragraph of explanation would have been good.

2. My exam results were almost perfectly inversely correlated with my TMA results. In other words I did worst on the topics which I knew more about! That was because the time pressure in the exam was extreme. Answers often required manual calculations that were conceptually straightforward but time-consuming (especially if you checked results).

The exam was testing ability to use a calculator fast as much as anything else. Result I ran out of time badly. The exam should either allot more time or lighten the manual computation load.

Course starting: **October 2014**

Review posted: **August 2015**

While the underlying principles of the topics taught in M249 are suitable for a level 2 OU module, the underlying mathematics is not. Consequently, the focus is on statistical ideas and methods, rather than mathematical technicalities. Theoretical mathematical statistics is covered in the level 3 statistics module M347 Mathematical Statistics.

We are sorry to hear that you had problems with time in the exam. This is very helpful for us to know and will help inform decisions concerning the format of future exams.

An interesting module. The title could do with changing as it's not as broad as it sounds. There are four topics on the course. You study 2 compulsory and choose a third from 2 optional topics.

The compulsory topics are Human Genetics and Health Issues and Empire of the Microbes. The optional topics are Frozen Planet and Molecules, Medicines and Drugs.

I choose Frozen Planet which was the best of the lot. Interesting, well written and makes use of BBC documentary footage. Human Genetics and Health Issues was interesting, particularly the sections on DNA and DNA mutations. I didn't enjoy Empire of the Microbes as I found the topic rather dry and boring. Due to new research several additions had to be added to a couple of the books but for the most part the books are engaging and full of interesting information.

Overall it's a fairly decent module. Very beneficial as a first module for students looking to do a science degree.

Peter Patching

Course starting: **October 2013**

Review posted: **February 2015**

An excellent fairly practical module which looks at the four statistical areas of: Medical Stats, Time Series, Multivariate Analysis and Bayesian stats.

The books are well presented and the information is fairly well written - personally I did get stuck on the Prosecutor's fallacy which I feel isn't written clearly in the book - but overall the course books are very good.

Aside from bits which are relevant to each topic area the maths is fairly limited to standard statistical analysis. While I would recommend taking M248 - just so you have the background knowledge as provided by the OU - a basic introductory book on statistics would be more than adequate. This is not a maths heavy course as it relies more on computer analysis - IBM SPSS - rather than a lot of pen-and-pencil calculations.

The TMAs are challenging but aren't overwhelming. They certainly help consolidate what you've learnt from the textbooks. The exam was of the CME format, where you have a question paper and an answer sheet and you choose the option (out of up to 8 potential answers) - it's a multiple choice exam basically.

Overall a module well worth taking, a lot of fun and very, very interesting.

Peter Patching

Course starting: **October 2013**

Review posted: **October 2014**

This is a great module. To be honest my interest is mathematics and not statistics. I only chose to do it because of the transitional arrangements and a need to get 60 credits this year (when I also studied MS221).

I was really surprised by how much I enjoyed it, all four blocks were well described and interesting. I would actually not hesitate in recommending this module to other Maths (or engineering) students.

Before the course I thought I'd learn more techniques to analyse data, which I did, but there was so much more to it than that and the words "Statistical Modelling" do not do justice to it.

For example you could apply what's taught in the time series block to extract very useful information about some data you have. If the plot is jagged and irregular you can work out a trend, seasonality, smooth out any random fluctuations and predict what results you would get next.

One final point, I had NOT done the module "Analysing Data" before hand, and did not find this put me at a disadvantage at all (I had done "Using Mathematics" which introduced Normal Distributions).

Paul Culliney

Course starting: **October 2012**

Review posted: **July 2013**

I struggled with this course because I did M248 alongside. I really wish I'd done M248 first so I had the basics in place. That said I'd got there by the exam & managed a grade 1 pass which I'm delighted with. Only just scraped enough on the TMAs & did much better in the exam than the TMAs. That seems to have happened to several others on this presentation.

I'd advise using the handbook as early in the course as possible. Block 2 was hard, apparently the hardest. Block 4 also troubled me but hey - I got the grade.

SPSS was heavily included in the TMAs, as was winbugs although I got through ignoring most of the computer chapters (not recommended). My main piece of advice would be to read the exam questions & take time to understand them.

Tutor issues marred the first blocks but were quickly resolved one I contacted the OU. In summary, not an easy course, but absolutely passable & contains useful stats content.

Course starting: **October 2012**

Review posted: **July 2013**

Course material - at the time that I took this course, it was separated into 4 blocks: Medical Statistics, Time Series, Multivariate Analysis and Bayesian Analysis. The material is very well presented and enjoyable to read (particularly Medical Statistics and Bayesian Analysis). The exercises and solutions are a must and an essential revision tool. The computer books accompany each block, and are also well written and very easy to follow. I would have liked to have seen more material on Bayesian Analysis (I suppose this is left for Level 3 courses) simply because this block is the best of them all! If you haven't come across Bayes Theorem yet in your studies, then this block will really open your eyes!

Course software (SPSS/WinBUGS) - SPSS is version 13 and somewhat out of date, but gets the job done. The computer books are very essential to work your way through the TMAs (more on this later). WinBUGS is an interesting addition, which compliments the Bayesian Analysis block very well.

Forums - these were amazing! When I found certain aspects difficult, I thought I was alone, however on entering the forums I was amazed at the camaraderie and help offered. Nearer the exam time, the solutions offered and posted really helped to focus revision to a fine art. I hope you have the same experiences with these forums as I did.

TMAs - all in all, I found these disappointing, not because they were difficult, but because they really were just variations on a coursework theme (thus finding the correct problem in the course material would allow you to complete the questions without any hassle).

The other annoyance was that they were heavily SPSS orientated, and pretty much all of the questions required output from SPSS to be included. This isn't that much of a problem (and indeed the programs are a necessary evil), but if you are like me, and tend to do these TMAs in your spare time, having to use a computer proved irritating when it had to wait until you got home after a days work etc.

However, yet again, the last TMA for Bayesian Analysis was excellent, I enjoyed this one the most and the WinBUGS output was kept to minimum (indeed, this is the only TMA that really made me stop and think about the answers).

Examination - past examination papers are available, the course offers very good revision exercises, so if you put the time in and notice the structure of the examination, then it will not prove a problem (trust me).

Summary - I obtained a distinction in this course, and I would happily take it again, even though the TMAs were frustrating. The course material, revision exercises, computer books, forums, access to past examination papers and solutions, and the sheer delight of the Bayesian Analysis block are priceless features that allow me to thoroughly recommend this course!

Course starting: **February 2011**

Review posted: **January 2012**

A highly enjoyable course - and I'm a professional statistician who took this course only for the Bayesian inference module. I found all the modules contained relevant, up-to-date and useful practial applications of statistics, without expecting a complete understanding of the mathematics underlying them. As a mathematical statistician, I was impressed that the explanations gave enough Maths to make the pedant like me interested, but not so much that those who needed to apply the techniques turned off. I recommend this to anyone needing to USE statistics, and also for those who need to understand its real-life uses.

Frances Mary Clare Drever

Course starting: **February 2009**

Review posted: **July 2010**

M249 is definitely my favourite course so far. The content was fairly straightforward to follow, the materials were excellent and my tutor certainly inspired me. I particularly liked the Bayesian Unit and the software to go with it was great to use. The approach I used was to push ahead with the course material at quite a strong pace and leave plenty of time at the end to revise for the exam. The exam is all about time management and being very familiar with the Handbook. Make plenty of notations in the Handbook to help yourself along. If you are interested in Statistics I would really recommend this course.

Mary Kelleher

Course starting: **February 2009**

Review posted: **January 2010**

This course is perhaps a little unusual in that rather than studying one topic, you study four separate branches of statistics in four blocks - medical stats, time series, multivariate analysis, and Bayesian stats. Personally, I found the course extremely interesting, with each individual block being very different from the others.

Each block is separate from the next so if you struggle with a concept in any one block, it should not affect your study of a different block. Each block is well-illustrated with plenty of real-world examples which help immensely, and are assisted with the course software. You need to use the course software extensively for the TMAs. I found Bayesian stats the hardest of the four blocks, and having seen the exam results breakdown, so did most people; nonetheless, the concepts are well expressed so that with perseverance it all becomes clear.

With M248 stated as a prerequisite, I was a little apprehensive about doing M248 and M249 together in the same year. The overlap between the two is primarily in medical statistics, with hypothesis testing being required early on. If you have already done M248, I would recommend reviewing hypothesis testing prior to starting M249. If you are doing M248 and M249 in the same year, you will find that some concepts in medical statistics may need a lot of concentration initially, however everything falls into place once block C of M248 has been completed.

The course software (SPSS) is both user-friendly and very powerful, although I couldn't get SPSS to run under Linux and had to resort to using Windows.

I found the exam very fair, and past papers and the specimen paper are good representations of what you could expect in the exam - do plenty of past papers and the exam will be no shock.

In summary I found this course very interesting, and with application a good course result can be achieved. I would recommend this course to anyone.

Course starting: **February 2009**

Review posted: **January 2010**

This is a very good course. The materials are clear and accurate and the tutor support excellent. I have a few complaints. These are more about what has not been covered rather than what is covered. I think that logistic regression could have been introduced in the medical statistics section. I also think that the more advanced Bayesian material did not go far enough. I could do all of the exercises in the course material and the TMAs but I do not feel confidence that I could use the WinBUGS software to handle my own problems even if these were of moderate complexity. I think that this material could be replaced with material that "leads on" from the earlier material on conditional probabilities (e.g. Bayes classifiers and Bayes networks) or revised. I also feel that some more links between modules would be useful. For example, the Bayesian module could have discussed screening tests as a link to medical statistics before discussing applications in other fields.

I must emphasise that these are minor niggles and I would recommend this course to anyone interested in refreshing knowledge gained from previous statistical courses, needing to "plug a gap" in their knowledge (my original reason for taking this course was to become familiar with a particular technique - I got far more out of it then I expected), or in gaining a grounding in modern statistics.

Course starting: **February 2009**

Review posted: **December 2009**

This is the best of the three OU statistics courses (M248,M249,M346) because it uses industry best-of-breed software (SPSS) and covers most of the ground of the other two whilst introducing important subject areas such as time-series and Bayesian updating. Like M248, the examination allows you to score 100% without getting everything right! You are allowed to attempt 120 marks but your actual mark is expressed as a percentage of 100 marks. This makes it easier, in my opinion, to get a distinction than on other Level 2 courses.

Andrew Rattray

Course starting: **February 2007**

Review posted: **August 2008**

A very disappointing course. The mathematical content is minimal and blurred by far too much use of a computer package. It seemed like a whole year's course on how to find the mean, median and mode in different guises.

If you want an easy course that requires little thinking and just using a computer package, this is the one for you. If, however, you like to understand the underlying mathematical princples then you won't find this a very interesting course.

The exam was very fair and similar to the sample paper. The TMAs were pretty straightforwad, albeit tedious.

Neil Reynolds

Course starting: **February 2007**

Review posted: **June 2008**

Whereas the underlying principles of the topics studied are suitable for a second level OU course in statistics, the underlying mathematics is not. Consequently the course focuses on understanding the underlying principles rather than on technicalities.

Throughout the course, the emphasis is on practical real-world applications.

For such applications, statistics is done using a computer and cannot be done in any other way. The use of statistical packages is therefore an essential element of the course, as it is to practicing statisticians in the real world.

Overall a great course. Good materials and a good exam. And I had a great tutor too which was a bonus. I really enjoyed it and would recommend it.

However, do it before the Level 3s (I couldn't as it wasn't available) as it will help you do better.

Course starting: **February 2007**

Review posted: **January 2008**

Each of the views expressed above is an individual's very particular response, largely unedited, and should be viewed with that in mind. Since modules are subject to regular updating, some of the issues identified may have already been addressed. In some instances the faculty may have provided a response to a comment. If you have a query about a particular module, please contact your Regional Centre.

To **send us reviews **on modules you have
studied with us, please click the sign in button below.