1. Patients launch a $1.27 million crowdfunding campaign for ME/CFS gut microbiome study.
    Check out the website, Facebook and Twitter. Join in donate and spread the word!
August 8th - What is the one thing about suffering with severe ME that the world needs to know?
Andrew Gladman brings our coverage of the Understanding & Remembrance Day for Severe ME, airing the voice of patients ...
Discuss the article on the Forums.

Free Coursera/Princeton online Statistics course starts this week

Discussion in 'Other Health News and Research' started by Dolphin, Sep 22, 2013.

  1. Dolphin

    Dolphin Senior Member

    Messages:
    6,519
    Likes:
    4,898
    A friend just alerted me to the following starting this week:

    https://www.coursera.org/course/stats1

    A couple of people, Simon and Graham, did this course (I think it was) last year and seem quite knowledgeable now.

    It would be good if there were more people with similar knowledge.

    Unfortunately, I don't think I should take it on due to ongoing commitments (I haven't done any coursera courses) but I have done some reading/revision in the past partly to try to help the cause.

    Conflict of Interest statement: I've no connection to the course. (It's free anyway so not sure whether any point me saying this)
  2. alex3619

    alex3619 Senior Member

    Messages:
    6,805
    Likes:
    10,324
    Logan, Queensland, Australia
    I would love to do this, but my math brain is in long-term hibernation. I would encourage anyone who wants to understand the science to do something similar.
  3. Graham

    Graham Senior Moment

    Messages:
    776
    Likes:
    1,842
    Sussex, UK
    It was a tough course at the start, and eased up as it went on. Parts of it had a vertical learning curve because key phrases in the use of the software had not been introduced. But it covered a number of important statistical concepts, and it was interesting to see how much emphasis the psychologist who taught it put on ensuring that data was well-behaved (in other words, not clumped or heavily skewed) before using the mean and variance to perform various checks. I did find that this one course put me pretty much at my energy limits for the first few weeks.

    Where it is described as "a comprehensive yet friendly introduction to statistics" it uses the same sort of understatement that my first year of university level maths labelled a course "Elementary Analysis". I used to think that with 12 years of studying maths at school, two A-levels in it, competing with so many to get a place, and I still am only at "elementary" level!

    Don't be put off by my comments though. I'm a mathematician, and statistics is an alien world to us. It's similar to the three main sciences: pure maths and mechanics are like physics and chemistry, but statistics is more akin to biology. It is actually quite unusual to find a mathematician who thrives on statistics. When I was at university in the late 60s, statistics wasn't even part of the mathematics faculty, and played no part at A-levels.

    If anyone does take it up, I'd be happy to help out where my memory has survived, and I'm sure Simon would do so as well. Mind you, I'm not sure how much has survived across my two brain cells. Simon's young and vibrant brain will of course be much more in tune.
    beaker, Dolphin and Simon like this.
  4. alex3619

    alex3619 Senior Member

    Messages:
    6,805
    Likes:
    10,324
    Logan, Queensland, Australia
    Philosopher's joke: Drinking alcohol causes loss of brain cells. As we lose cells, the remaining cells become more and more connected. So if you drink enough you lose more and more brain cells until at last only one cell remains, but its infinitely connected. You have become God!

    So you are only one cell away from perfection, Graham . ;)
  5. physicsstudent13

    physicsstudent13 Senior Member

    Messages:
    225
    Likes:
    5
    I'm studying math at university and was wondering if there any good nootropics like acetyl carnitine, citicoline, magtein?
  6. Valentijn

    Valentijn Activity Level: 3

    Messages:
    5,869
    Likes:
    7,936
    Amersfoort, Netherlands
    Cool, I've just signed up. It starts today, and the first due date for completing an assignment (quiz) is on the 29th. I think the main page says it takes 5-8 hours per week to view all the info, etc.
    rosie26, Dolphin, Simon and 1 other person like this.
  7. Simon

    Simon

    Messages:
    1,239
    Likes:
    3,694
    Monmouth, UK
    Good luck Valentijn

    If Graham thinks I'm young, then he's a lot older than I realised...
    I think it's fair to say that my ability to talk convincingly about stats improved dramatically after this course, not sure about the knowledge, though.

    It is a lot of work, and hard going, but the worst thing about the course I did was they often inadvertently didn't give you the all info you needed to do the assignments, but I know they were spending a lot of time ironing out the wrinkles. And when they say 'fun' in the course prosepectus, they are lying. However, well-taught and a very good course - just don't expect to be able to do anything else in your life for the duration.
  8. Dolphin

    Dolphin Senior Member

    Messages:
    6,519
    Likes:
    4,898
    Perhaps somebody can help you on Phoenix Rising, but seems a bit off-topic for this thread?
  9. Valentijn

    Valentijn Activity Level: 3

    Messages:
    5,869
    Likes:
    7,936
    Amersfoort, Netherlands
    My main interest is just in being able to easily read the statistical data in research, though it'll be a bonus if I understand it well enough to detect problems.

    Some things I've figured out the hard way (p values, beta effect sizes, odds ratios, etc), but it can still be a slow struggle to find the specific bit of data I'm looking for. That is, is the effect of an SNP significant, which allele is being looked at, is it making things worse or better, and how much worse or better. Usually I can figure these things out, but there's still too many times where I can't :p
    beaker, rosie26, Bob and 1 other person like this.
  10. Hope123

    Hope123 Senior Member

    Messages:
    1,141
    Likes:
    531
    I've taken stats before and wanted to add that contrary to what some might think, basic-intermediate stats is NOT really math heavy. That is, if you know algebra, you can do stats besides which computer programs do a lot of the heavy lifting with analysis. It's mostly about understanding concepts.

    My suggestion to those who don't know if it'll work for them is just sign up; there is no penalty or obligation to watch any lecture nor to complete any assignments. They also tell you which weeks cover what topic so you can just tune in on the ones that you're interested in.

    In my view, stats is so much more useful than the year of calculus I had to take for college. I hardly use calculus ever after freshman college but stats, I still use. [And it's useful not only for science but for business, the humantities, engineering, social sciences, anthropology, etc.]
    rosie26 likes this.
  11. user9876

    user9876 Senior Member

    Messages:
    684
    Likes:
    1,549
    I've signed up as well. Some bits of stats I know well but other pieces I've no idea about so I thought it would fill in gaps. I've also been intending to learn R for some time.
    Simon, rosie26 and Bob like this.
  12. Graham

    Graham Senior Moment

    Messages:
    776
    Likes:
    1,842
    Sussex, UK
    Ah, but I've just stopped drinking alcohol and coffee. Nothing to do with the ME. It's my age and the prostate. I'm struggling to get round a complete walk with the dog without breaking into a run when I near home.

    And Hope123, I do so agree with you. When I was teaching, we decided to put a lot of emphasis on teaching statistics properly at my school. In my opinion people shoot on to the complex stuff too quickly, before they have even understood why we have three different categories of average, what measures of spread represent, or, most tellingly of all, how percentages work and how they deceive (Bob will confirm what a tedious nit-picker I am at that!). A good knowledge of stats doesn't need flashy maths, and it is so much more important these days. Let's face it, the errors on the PACE trial were nothing to do with sophisticated statistics.

    Valentijn, the course doesn't really hit the level you mention. Instead it puts a much more thorough understanding on the layer below that (or that's how I see it) and it forces you to think with an extra layer of clarity. I'd come across a chunk of the stuff before, and even taught some of it, but I had never been forced to think so carefully about what was going on. For that reason alone, it is an important course (and should have been compulsory for the authors of the PACE trial). But it was great having Simon do the course at the same time so that we could discuss stuff and sort out our ideas.

    Physicsstudent13: I'm an old wrinkly, and ME has wrecked my health and knocked chunks off my mental performance. I can only dream of returning to my old, original pure brain power, and nothing on earth would tempt me to take risks with it. Sorry if that's a bit of a patronising reply, but I now value good health way way over anything else, and of all the limitations of ME, the brainfog is the part that I hate the most. If you have ME yourself, then I can only admire you for getting that far!
    Valentijn likes this.
  13. Simon

    Simon

    Messages:
    1,239
    Likes:
    3,694
    Monmouth, UK
    The first in aseries of fairly basic, short and accessible articles on statistics:
    Points of significance: Importance of being uncertain : Nature Methods :

    Normally, the Nature Methods journal focuses on new high-tech methods, but they obviously feel that a lot of scientists need a bit more help with stats.

    This first column includes what to do in a Game Show when you first pick one of 3 doors, and a prize is only behind one of them. The show Host then opens one of the 2 doors you didn't choose to reveal no prize - and asks if you want to switch doors. Statisitcs gives a clear answer of what to do. See, statistics is interesting and relevant. Really it is. Hello, is anyone still there?
    Valentijn likes this.
  14. user9876

    user9876 Senior Member

    Messages:
    684
    Likes:
    1,549
    If its like the statistics used in certain medical trials the probability of getting the prize is 0 as they will switch it from what ever you choose.
    Valentijn likes this.
  15. Valentijn

    Valentijn Activity Level: 3

    Messages:
    5,869
    Likes:
    7,936
    Amersfoort, Netherlands
    The have a show in the Netherlands called the "Postcode Lottery". After some silly stuff at the beginning, they go onto a game which has become known as "Deal or No Deal" in various countries around the world.

    Basically they have a couple dozen closed suitcases (held by scantily clad women, of course :rolleyes: ) each with an amount of money written inside, ranging from one cent to 1,000,000 dollars or euros. The contestant chooses one suitcase, then selects several other suitcases to be opened in batches. After each batch, the "banker" makes an offer based on what might be in the contestant's chosen suitcase, which is recalculated based on what is known to not be in the suitcase.

    Thus after the contestant selects several suitcases, the offer will go up if low values are revealed (greater odds that there's more money in his or her suitcase), or go down if high numbers are revealed. So really, it never actually matters which suitcase the contestant picks, unless he or she is an idiot and insists on opening every suitcase on the stage, until it's known what's in the contestant's suitcase. What really matters is opening enough low value suitcases, to get a nice and high offer from the "banker".

    So this should be a matter of luck, aside from knowing when to stop. And that's what happened every time I watched it in the US. But the first time I watched it in the Netherlands, the contestant who won some earlier rounds (trivia type stuff), had a bit of a surprise in store.

    He had been watching the show closely for a very long time (probably years), and had come up with a system: he had observed that the same suitcase basically never have the big amounts in them, if they've had those big amounts in them before. Hence idiot contestants who think a number might be "lucky" because it held a big amount in the past would be likely to get 10 cents, or a dollar, etc.

    So he started reading from his list, the numbers of the suitcases which had recently held big amounts, and they were opened up, one at a time. Before long, he had an offer of 300,000 euros or so, and stopped, even though his system seemed to be working wonderfully and he only opened up one or two with high values, and most high values were still on the board, and thus possibly in his suitcase.

    If he'd kept playing (they had him keep reading his list, just to see), he would have had over 500,000 euros, but his suitcase ultimately had something less, if I recall, along the lines of 100,000 or less? Hence the host was inflicting a strong case of the "what-ifs", which are great for trapping the suckers that want to keep pushing, because they'd rather lose something that they never really "owned" than miss out on the opportunity to get a higher amount.

    My thought was that it was much like playing a game of poker - as long as a good system is being used, and the player is sticking to it, they're making the "right" moves, even if they lose. The actual outcome on a specific game didn't matter as much as using the right system does in the long run.

    I haven't watched the Dutch version of the show since that one remarkable game, but I'm willing to bet they've abandoned their old system of putting low amounts in the suitcases which recently held the big money.
    Dolphin, Simon and rosie26 like this.
  16. rosie26

    rosie26 Senior Member

    Messages:
    957
    Likes:
    1,126
    NZ
    I used to watch that game a lot. It used to amuse me when people who called themselves psychic had a go, I don't think any of them one the big one :)
    Valentijn likes this.
  17. biophile

    biophile Places I'd rather be.

    Messages:
    1,350
    Likes:
    4,023
    Statistics One looks fairly comprehensive for an introductory course, but 5-8 hours per week for 12 weeks may be too difficult to guarantee for many patients depending on health and life events. I would not mind doing this if I could, but right now I have a few other things going on which are either more important or more rewarding. Maybe next time if possible.

    For those who are more interested in brushing up on the basics for interpreting research papers, this simpler course may be more relevant or easier, although some goodies are missing:

    https://www.coursera.org/course/statreasoning

    Also consider the following two books which may be of some help:

    - How to read a paper. (Greenhalgh).

    - How to lie with statistics. (Huff). ... an old classic.
    WillowJ and Dolphin like this.
  18. physicsstudent13

    physicsstudent13 Senior Member

    Messages:
    225
    Likes:
    5
    Dear Graham, my mathematician friend teaches statistics at UW Madison, yes I have terrible brain fog from sleep apnea also and perhaps NAFLD- so ME is a swelling of the brain? I started to have less fog after starting NAC, phosphitadyl choline, but am exhausted on this asv ventilator
  19. Snow Leopard

    Snow Leopard Senior Member

    Messages:
    2,218
    Likes:
    1,568
    Australia
    Does this course teach you how to do statistical data processing with any sort of professional software, eg SPSS etc?
  20. Simon

    Simon

    Messages:
    1,239
    Likes:
    3,694
    Monmouth, UK
    Yes, and it's free.

    The bad news is it's R, a command line programme, which has a steep learning curve. However, R is maintained and updated by a worldwide community of academics (practising statisticians) so is reckoned to be very good indeed. all the updates/plugins that ever found were hosted on university sites. R is also often mentioned as the core tool in academic publications. And slightly oddly, given it's a command line tool, it's reckoned to have the very best graphics of any package.

    Oh, and it's based on a statistical language 'S' (clever, eh) that I think is used to underpin commercial packages. But the developers of S went on to build R.

    Valentijn likes this.

See more popular forum discussions.

Share This Page