Thursday, August 2, 2012
The problem with time off...
I have just returned from Improving University Teaching in Innsbruck, a small, but friendly international conference. I had a great time, met new people, reconnected with old friends, but as far as Stats101 goes, I did nothing. So now I'm feeling like one of the students who hasn't kept up with the work and is now panicking about the end of course test. That's an interesting feeling. I received an email about the end of course test, and how some people had already done it (!), and that there would still be time to complete it, in order to receive the completion certificate (which, I have to admit, I didn't realise was part of the deal.). So, now I have a challenge - can I remember what I was doing two weeks ago, so that I can finish the course?
Friday, July 13, 2012
All MOOC'd out
I haven't had much time to do any Stats 101 this week, but yesterday I found myself with a bit of time. I logged into the course, and immediately had a Homer Simpson moment. Don't leave the course in the middle of a section. D'oh! It seems obvious, but the last time I was working on it, I got to a quite important explanation and left it before doing the examples (normalization, posterior probabilities) so when I returned to the course, I looked at the screen with what I imagine was the same look that students often have when we ask them something they obviously don't know. I just could not get my head around the example in front of me, so I had to go back, and look at the previous video again, and then I was aware that I was copying the calculations and plugging in the numbers, but I really didn't understand what I was doing, and I was pretty confident that if the prompts were taken away and I was asked to perform the calculation, I wouldn't be able to do it. Mmhmm. Something to think about...
Coincidentally, earlier on I had looked at Ray Land's keynote at the Threshold Concepts conference in Dublin. He was talking about students' understanding (or lack of it) and how they go through a process of mimicking - using the language and the conventions of a subject or a discipline without really understanding what they are, but which is probably a powerful step on the road to understanding. This is something that we, as educators, need to be on the lookout for, because sometimes confident use of the language can mask lack of understanding at a deep level. I wonder if our examinations system encourages this to a certain extent - MCQs with verbal prompts in the answers as opposed to writing on a blank piece of paper, or researching a topic for a presentation or...but I digress. (I could start to talk about Bloom's (or Graham Gibbs' SOLO) Taxonomy here, but I won't.)
Yup, I was definitely mimicking all right. Ray also talked about psychological capital, and how a combination of factors is required for someone to succeed - a combination of optimism, hope and perseverance (the references to the literature are in Ray's talk, so please watch it.). I'm not put off by my hiccup on the Stats 101 course, in fact, I predicted it in the first post (maybe it happened because I predicted it?!). My motivations for doing the course are to understand statistics in order to complete a doctorate, so I'm pretty motivated! (And I can get help from AQMen if I need it, but I feel that I have to understand it well enough to be able to ask the right questions.) I did move on and completed the other examples, and managed to do them (mimicking all the way), and have moved on to programming Bayes Rule (Anyone doing Udacity ST101 - indent the program otherwise it returns an error). I am still confident that I'll keep going on with the course, and I'm still on the lookout for that point when I really start to understand it...
Coincidentally, earlier on I had looked at Ray Land's keynote at the Threshold Concepts conference in Dublin. He was talking about students' understanding (or lack of it) and how they go through a process of mimicking - using the language and the conventions of a subject or a discipline without really understanding what they are, but which is probably a powerful step on the road to understanding. This is something that we, as educators, need to be on the lookout for, because sometimes confident use of the language can mask lack of understanding at a deep level. I wonder if our examinations system encourages this to a certain extent - MCQs with verbal prompts in the answers as opposed to writing on a blank piece of paper, or researching a topic for a presentation or...but I digress. (I could start to talk about Bloom's (or Graham Gibbs' SOLO) Taxonomy here, but I won't.)
Yup, I was definitely mimicking all right. Ray also talked about psychological capital, and how a combination of factors is required for someone to succeed - a combination of optimism, hope and perseverance (the references to the literature are in Ray's talk, so please watch it.). I'm not put off by my hiccup on the Stats 101 course, in fact, I predicted it in the first post (maybe it happened because I predicted it?!). My motivations for doing the course are to understand statistics in order to complete a doctorate, so I'm pretty motivated! (And I can get help from AQMen if I need it, but I feel that I have to understand it well enough to be able to ask the right questions.) I did move on and completed the other examples, and managed to do them (mimicking all the way), and have moved on to programming Bayes Rule (Anyone doing Udacity ST101 - indent the program otherwise it returns an error). I am still confident that I'll keep going on with the course, and I'm still on the lookout for that point when I really start to understand it...
Monday, July 9, 2012
Week 1, in which I think about MOOCs
I am taking part in a MOOC. Actually I'm not sure that that's the correct grammar. What does one do with a MOOC? Let me begin. Massive Open Online Course. This one in particular is the Udacity Statistics ST101 course. Why am I doing it? Well, I am currently at the end of the second year of a doctorate in education. I am at the point where I am ready to start collecting data, and, never being one to take the easy option, I have decided that I have to collect quantitative survey data before I delve into qualitative data. I have one problem, though. I have carefully managed to avoid any intimacy with statistics up until this point (with the exception of Chi squared which we do with our first year students, and with about twenty years' of practice, I'm quite good at).
So, why a MOOC? I have several possible options. I could sign up for the first year statistics course at my home institution, pay my part time fee, and sit with my first (and second) year undergraduate students. There are some problems with that. I teach 10am-1pm and/or 2-5pm and statistics lectures are at 1pm. That's a LONG day. And I really don't want to be sitting in a tutorial with my students - not because I don't like them, but I wonder what my being there as a student would do to them. I could go to a postgraduate training course at my home institution or at my doctorate institution, but that again would conflict with teaching time. So, I opted for a MOOC, which I can do online, at my own pace, in my own time. What could possibly go wrong?
I have experience of online and distance education. I completed a Masters module in Enquiry Based Learning at the Universities of Strathclyde and Helsinki. This was based around Moodle discussion and we all pitched in from the UK, Nordic countries and Russia. I enjoyed it a lot, drawing from the experiences of the other people on the course, and completing the assessed coursework from my own experiences. I think that the most important part of this course was the peer support - we "talked" with each other all the time, supported by our tutors. I also completed a certificate program in SoTL Leadership at the University of British Columbia. This was a strange (not in a bad way!) one because I was the only one of my cohort who was not doing the course on site. I pretty much completed the course on my own apart from two exercises: one peer feedback exercise, and one collaborative exercise, where I was so excited to be talking to my classmates that I stayed up all night emailing them about our project. Again, the human contact, with my tutor, and with the peer group, was very important to me.
Now, the MOOC. I have been doing it for one week, and so far things are going well. I have progressed from graphs, visualizing data, probability & conditional probability to section 10: Bayes Rule in a week. I'm currently working my way through the examples. So far, I got all the questions right apart from one right at the beginning and a couple where I had to take a second go. I can hear you asking, "What kind of course is that then? It sounds far too easy." I am still at an early stage in the course, and I do have in the back of my mind that at some point I am going to hit a wall of unintelligible numbers, but it hasn't happened yet. Let me try and describe the way the course works.
The course is taught by Sebastian Thrun, a computer science professor at Stanford. Through a series of YouTube-type videos, Sebastian takes a step-by-step approach to solving statistics problems. Literally. His videos never last more than a few minutes, and each example is carefully explained, and followed by an opportunity for the student to complete a calculation or make a prediction, which can then be entered on the video interface. You're instantly told if the answer is correct, and if it's not, you can try again (I don't know if you get an infinite number of gos!). I like this approach. It's nice and slow, and for me, with no background in statistics, I can keep up with the explanations. As I said, I'm up to section 10 in a week, and it does occur to me that there's something of the learning equivalent of boiling frogs in this approach; that the heat is being turned up slowly and at the moment it's too early to discover if I'll turn out to be a statistics genius or be boiled alive.
I was initially slightly concerned by the pace of the course - it really is ONE-STEP-AT-A-TIME, and this was reflected in the forum comments. I've since revised my opinion on that. I like the stepwise approach, and I've stopped looking at the forum comments. However (and I realise I'm at the beginning of the course) I do wonder about being able to take what I have learned on the course and applying it to my own data. How that will transfer, I don't know, and I guess it's too early to make a prediction.
One thing that did surprise me was that I don't feel that I need peer contact. This surprised me a lot, both from my previous experience with online learning, and reading John Naisbett's material. All I can say is that when Sebastian Thrun is talking me through the problems, it feels like we're having a tutorial conversation. I haven't hit the wall yet. I might feel different when I do.
So, why a MOOC? I have several possible options. I could sign up for the first year statistics course at my home institution, pay my part time fee, and sit with my first (and second) year undergraduate students. There are some problems with that. I teach 10am-1pm and/or 2-5pm and statistics lectures are at 1pm. That's a LONG day. And I really don't want to be sitting in a tutorial with my students - not because I don't like them, but I wonder what my being there as a student would do to them. I could go to a postgraduate training course at my home institution or at my doctorate institution, but that again would conflict with teaching time. So, I opted for a MOOC, which I can do online, at my own pace, in my own time. What could possibly go wrong?
I have experience of online and distance education. I completed a Masters module in Enquiry Based Learning at the Universities of Strathclyde and Helsinki. This was based around Moodle discussion and we all pitched in from the UK, Nordic countries and Russia. I enjoyed it a lot, drawing from the experiences of the other people on the course, and completing the assessed coursework from my own experiences. I think that the most important part of this course was the peer support - we "talked" with each other all the time, supported by our tutors. I also completed a certificate program in SoTL Leadership at the University of British Columbia. This was a strange (not in a bad way!) one because I was the only one of my cohort who was not doing the course on site. I pretty much completed the course on my own apart from two exercises: one peer feedback exercise, and one collaborative exercise, where I was so excited to be talking to my classmates that I stayed up all night emailing them about our project. Again, the human contact, with my tutor, and with the peer group, was very important to me.
Now, the MOOC. I have been doing it for one week, and so far things are going well. I have progressed from graphs, visualizing data, probability & conditional probability to section 10: Bayes Rule in a week. I'm currently working my way through the examples. So far, I got all the questions right apart from one right at the beginning and a couple where I had to take a second go. I can hear you asking, "What kind of course is that then? It sounds far too easy." I am still at an early stage in the course, and I do have in the back of my mind that at some point I am going to hit a wall of unintelligible numbers, but it hasn't happened yet. Let me try and describe the way the course works.
The course is taught by Sebastian Thrun, a computer science professor at Stanford. Through a series of YouTube-type videos, Sebastian takes a step-by-step approach to solving statistics problems. Literally. His videos never last more than a few minutes, and each example is carefully explained, and followed by an opportunity for the student to complete a calculation or make a prediction, which can then be entered on the video interface. You're instantly told if the answer is correct, and if it's not, you can try again (I don't know if you get an infinite number of gos!). I like this approach. It's nice and slow, and for me, with no background in statistics, I can keep up with the explanations. As I said, I'm up to section 10 in a week, and it does occur to me that there's something of the learning equivalent of boiling frogs in this approach; that the heat is being turned up slowly and at the moment it's too early to discover if I'll turn out to be a statistics genius or be boiled alive.
I was initially slightly concerned by the pace of the course - it really is ONE-STEP-AT-A-TIME, and this was reflected in the forum comments. I've since revised my opinion on that. I like the stepwise approach, and I've stopped looking at the forum comments. However (and I realise I'm at the beginning of the course) I do wonder about being able to take what I have learned on the course and applying it to my own data. How that will transfer, I don't know, and I guess it's too early to make a prediction.
One thing that did surprise me was that I don't feel that I need peer contact. This surprised me a lot, both from my previous experience with online learning, and reading John Naisbett's material. All I can say is that when Sebastian Thrun is talking me through the problems, it feels like we're having a tutorial conversation. I haven't hit the wall yet. I might feel different when I do.
Subscribe to:
Posts (Atom)