Observation exercise
Mood:  a-ok
Now Playing: Wooden Heart (Elvis Presley)
Topic: Methodology
So, went to Cranfield on Monday and did the observation exercise - I'm not quite sure what is it so new that I learn or whether that helps me to formulate the research question any better ... I'm really not sure f my future observation exercises might be pointless. Anyway, I'm supposed to do some interviews with some students later this week over telephone - so hoping I can find something that is better there.

Now let me see what I thought was unique or different about this observation ... well, there was:
1) A lot of concentration on the formulation of problems (so guessing this is a soft-applied context) and the first instance when the formulation problem was given, this was definitely not a mechanical problem - but more a constructive or perhaps interpretive problem - students had to figure out how to solve the problem. However, in fact this was not a formulation problem when one thinks about it, in fact it was a mixture of a formulation and solution problem, because students had to think about what constraints could affect the answer - and guess an answer in the end.

Now, looking back and reading the quotes I've written it is getting a lot more interesting - probably I didn't think through all I could do with this data.

Questions I want to pilot
Mood:  a-ok
Now Playing: That's What Daddy Wants (Wayne Hancock)
Topic: Research questions
So, got the observation set up over in Cranfield on Monday - and I am still wrestling with what I expect to get out of it.

So, I'm brainstorming here - all I keep remembering is that I'm looking at the learning and teaching of linear programming in various learning contexts - although it is really learning rather than teaching.

Ok, let's see what are the main goals of this pilot study:
1. Pilot my observation skills
2. Pilot my interviewing skills

So, essentially I'm piloting two methods of data collection.

Just got an idea - I wanted to use Galbraith and Haines mathematics computing attitude scales - I was thinking I could have pilot here in the OU with people who would have perhaps did linear programming or mathematics at the undergraduate level or perhaps using OU students - not sure what I'm looking for in that case - but just as pilot to compare how these attitudes may differ from discipline or possibly by students taking the course, if whether their attitudes of mathematics-computing compare well to the course that they were undertaking.

Or perhaps I can interview people who did linear programming here in the OU and as them how they felt about it, what kind of software they used and what they thought about the software etc. It might be a long time ago for these people and they mightn't even remember the course.

I could possibly interview OU students who did the course?? Got to go through SRPP for this - but still need to come up with what I'm looking for ... probably could call it exploratory research just like exploratory surgery :D

Still haven't got to the questions I want to get from the pilot study ... I think from the observation/ interview study at Cranfield I want to see:
1. What kind of approach is used in teaching linear programming for this course? I.e. is it more like a soft-applied approach or a hard-applied approach etc
2. How are the lectures presented?
3. The amount time spent on formulation, solutions etc and the problems given to the students
4. The amount of interaction that the students have between them and the teacher
5. When is the software whipped out for the students to use and what are the problems and comments that the students have when using the software
6. What they do when they are behind the computer - i.e. follow teacher's instructions to the law or do their own thing


From the teacher interviews:
1. Why the particular approach is used for these particular kind of students? i.e. soft-applied/ hard-applied - and whether a different approach would have been better in the teacher's opinion
2. Would this course be sufficient to meet their future goals in their field?
3. How much they think linear programming would be used in their future jobs etc.
4. Why the choice of software?
5. Would they have preferred a different software?
6. What problems they see with using the software and problems that student's bring up?
7. How do they go about assessing student's linear programming accomplishment
8. What kind of problems are set for them - in assignments, tests or projects?

As for the interviewing of students at Cranfield and maybe in the OU:
1. What they think of linear programming - is it a hard or easy concept?
2. Do they think it will be easy to implement? Or do they see themselves using the concept or using some other concept
3. What they found hard about linear programming?
4. What they enjoy most or least about studying it?
5. Do they enjoy using the software for linear programming? What they enjoyed the most/ least?
6. Would they have preferred some different kind of software?
7. What problems they see when using the software and learning?

Well, these seem all like questions I want them to understand - but to what purpose? To merely illustrate how linear programming is taught at one institution (or learning context) and to highlight the problems/ challenges of teaching linear programming (and with software)?

Seems good to me :D

Observation Study on the 6th
Mood:  a-ok
Now Playing: Listen to your heart (Roxette)
Topic: Methodology
I've set up an observation study of linear programming at Cranfield university (I think the guy over there might be a bit unwilling - but got the appointment so nevertheless - I thought I might be coming over as pushy - but Gill says for our PhDs we should be persistent).

Anyway, from the case studies workshop I attended was thinking I should take out some pictures - but wonder what are the ethical and legal considerations of that? Do I need to get the people to sign a photo release? Or the university's permission? I'm not even quite sure if I could just go and observe the students with the lecturer's say so without getting the university's permission!!

This is getting quite complicated ... but ... anyway, I'm doing it ... if it becomes a complication, probably can just say sorry :) and take my data and leave? Anyway, will have to write to the guy and ask ... will do it on Monday. Or maybe I should do it now? I don't know if that will be too pushy? But Gill said to be persistent .... hmmm ... bit ambivalent.

Well, I just created the email, all I need to do now is to send it but I wanted to tell him that I will like to talk to him about 10-15 mins before or after the class but I should at least tell him on what - but I have no idea what the what is as yet!!!

U500 Case Studies Worshop
Mood:  a-ok
Now Playing: I Swear (John Michael Montgomery)
Topic: Seminars
So, went to the case studies workshop on Tuesday, I attended the same one last year, but last year I had an entirely different focus ... whilst this year I have no focus :D. No that's a lie ... I do have some kind of focus, I know I'm looking at how linear programming is taught and learnt in different learning contexts. What I am not certain about is what these learning contexts are.

Anyway, I'm not sure if I'm going to employ case studies ... or if I do, it will just be some aspect of it rather than a full blown case study. The guy who did the workshop (Terry Newholm) in his presentation said that (I think he was quoting someone else) is that a case study is about 'researching an organism in its environment'. Now, that's an interesting concept but I think for the case of learning and teaching linear programming there really isn't any contained environment, as in a classroom, where learning only goes on - not sure what I'm getting at, but I think what I'm trying to say it is not a closed system but an open system (just to borrow some terms from thermodynamics), and I guess the 'total energy' for learning is not confined to the classroom :D (wonder if that even makes sense!).

Now a question, that Terry asked us to consider in our research what kind of cases and how many cases should we choose. Well, I obviously had to go from a modernist/positivist approach because it is the only approach I know that I'm comfortable with. So, using that approach, I decided I'll have three learning contexts ... and these learning contexts will be based on the disciplines or should I say focus of the course. That is ones that are hard-applied, soft-applied and hard-pure ... where:

1. soft-applied means the course is application intensive such that there is a high attention on formulation, on just getting the answer (i.e. no theory) and possibly some attention on applying the sensitivity analysis - this will perhaps mean that soft-applied courses will be most likely in using linear programming software - checking this theory against what I found in my MSc literature it shows that 92 % of the soft-applied did use software, but pure-applied were quite close to this also (90 %).

2. hard-pure means the course is means that there is a high attention on the theory of solving the problem and theory of finding the values for sensitivity analysis, possibly of using a wide range of mathematical notation and approaches such as algebraic, matrices, Lagrange etc.

3. hard-applied means the course places almost equal importance on the formulation, solving of the problems and solving the sensitivity analysis, through calculations and with a means for application, basically a very balanced course.

I wonder if I ask the students from each of these types of courses if they will prefer another learning context, for example if the soft-applied will prefer the hard-applied approach and whether this is dependent on the disciplines they come from or is in fact intrinsic to the their way of study (this will mean that I'm throwing out the theory that approaches of study are dependent on the course the student is studying for but rather intrinsic to the student) or whether this is matter of what the student expect to get from the course i.e. to pass the course, apply this in a job or understand a subject thoroughly.

Anyway, those will be my three learning contexts, but from each of them I can choose case studies from two more situations, one in which students have high computer affinity or ability (hopefully measured by an inventory) and have a low mathematics aptitude (again measured by an inventory) and then have cases chosen from the situation vice versa i.e. high mathematics ability and low computer affinity - that way I am able to get the two ends of the spectrum. However, from each of these contexts I should have at least have 2 or 3 cases. If I do that, that'll mean either 12 or 18 cases in all. Terry recommended have less than 12 cases to allow the reader to follow all the cases!

Anyway, even with all these cases I think I'm forgetting that I am trying to look at the benefits of intermediate steps or different types of softwares in providing the steps - I'm not quite sure how these cases will provide this, this will obviously be dependent on the type of software that is being used by each of the students - and how am I going to measure these benefits? Is it the benefits of learning linear programming with/without steps directly related to their , to their job prospects, to their understanding of the subject or to their passing of the job? How do I measure the benefit?? Hmmm ... that reminds me that Rebecca recommended a paper by Conole and Dyke to me on the concept of affordances to me which might help me in deciding what the benefits are - since they looked at the affordances of ICT - which I don't quite understand as yet - since I haven't read the paper - but I think it has something to do with the extent to what the software can do and the extent that people utilise this extent - or something like that - that's the gist I got from Rebecca.

My draft research question
Mood:  a-ok
Now Playing: Hippy (Atomic Kitten)
Topic: Research questions
So, based on my supervisory meeting last month in December, the research question that I'm currently working with is:

How do students learn linear programming using software in various learning contexts?"

I originally had learning environments but Doug thought learning contexts might be more suitable. As it is, this research question is still broad and they did suggest using a number of sub-questions to focus into what I need to research.

I'm back from vacation
Mood:  a-ok
Now Playing: Can't Help Falling In Love (Elvis Presley)
So, here I'm back from vacation and I feel a bit backward since everyone has sort of getting on with their research and I'm still in the same place I was last month.

Gill is starting her pilot this week!! And I haven't even got around to thinking about my pilot. I've all these tasks I had set up on MS Project and I haven't even looked at Project so it is likely that I'm falling behind.

I know that my pilot has to do with observations or something like that but I still have to get together a list of universities etc and likely lecturers and see when they are willing to let me come observe them. I think I'll have to do that today, so, I can feel as if I'm getting some stuff done :D.

Passed U800
Mood:  bright
Now Playing: Rainbow Rider (Tanya Tucker)
Topic: Thesis writing
Not sure where this topic fits in ... but hey, just got results for my dissertation got 82 ... passed with merit ... so that completes my masters!!! I hence automatically upgrade to Ph.D (never mind that we've been doing Ph.D work since October) - but now we know the work we've been doing since October is no longer in vain :D.

M373 and BM240 course materials
Mood:  a-ok
Now Playing: I Get Around (Beach Boys)
Topic: Data Collection
So, was looking at the BM240 (business students) and M373 (maths students) materials after the U500 seminar on epistemological awareness, and noticed quite a lot of difference from what is taught in BM240 and M373. Well first of all ... BM240 only has one section (or book) on it ... whilst M373 has 4 books (granted they teach a lot more).

In BM240 the students are guided by the "Quantitative Analysis for Management" 7th Ed. textbook by Render and Stair (2000), students are taught the formulation of linear programming problems in particular these kind of problems:
1. Marketing applications: media selection and marketing research
2. Manufacturing applications: production mix and production scheduling
3. Financial applications: portfolio selection

There were a lot more problems in the books like employee scheduling but these were not covered by the course.

Students were also taught the graphical method, but were not taught to solve it using the isoprofit line but rather using the corner-point solution method. Further, in the corner-point solution method they were taught to solve for the point using simultaneous equations rather than reading off the point on the graph. Students were not taught the simplex algorithm.

With respect to software, students used Excel solver, the problems were first formulated by hand and were then taught how to set up the problem in Excel to solve.

They were also taught the types of linear programming problems they might encounter:
1. Infeasibility (no solution!)
2. Unboundedness (no bound on solution - can increase as much as it wants!)
3. Redundancy (repeated constraints)
4. Alternate optimal solutions

With respect to sensitivity analysis, they were taught sensitivity analysis as a concept, but did not look at a printout of a sensitivity analysis, but rather solved problems using Excel with different values on the RHS and coefficients (I think) to see how this influenced the solution.

With respect to M373, they were also taught the formulation, but first they were taught it in the algebraic format and then to convert it into the matrix format. They were taught several types of problems including:
1. Diet
2. Production Process models(I think it was an example of a waste disposal)
3. Multi-period models
4. Blending

The graphical method was also taught, and the concept of feasible regions were looked at in detail. Also, they were taught how to solve problems using the optimal-point solution method (note how this term differs from corner-point solution although it is the same thing), in this case students had to read out the answer (I think) rather than solve with simultaneous equations.

Students were taught the simplex method, with a mixture of algebraic and matrix notation (looked really hard to understand!!!) - which they called putting it in its canonical form. They were also taught to solve it using the matrix format (which looked a lot easier to understand - wonder if its deliberate!)

They were also made to understand the concept of dual and duality theory. Hence from this solve the problem using the 2 phase simplex method (using the matrix notation) and hence considered the terms of excess and surplus variables. With using the duality theory, the concept of the sensitivity analysis was broached (which they called post-optimality analysis) and were able to calculate the shadow price and reduced cost through the use of the matrix notation. I can't remember if they looked at the range/ limits of the RHS and coefficients of the variable. Will have to double check.

They were solving problems using the computer but I'm not quite certain how they did it since the computer activity book was not included in the materials in the library. I've written to M373 course manager to see if I can get a copy - hopefully will get a reply tomorrow and then can update this accordingly.

I have no clue about research questions!!!
Mood:  don't ask
Now Playing: Jack The Ripper (Link Wray)
Topic: Research questions
Well, I'm suppose to make an entry into my research blog today .. but have no clue what to write about since I'm pressed for time (got to go catch a bus).

I've read so many papers and yet I'm clueless about my research question. I think I'm just going to use 'How do students learning linear programming using software?" and then put a lot of sub-questions to answer that and make up a methodology of observations, interview, questionnaires (inventories) and examination of scripts. Although it seems like a lot of work and I have no point to why I'm doing it and how it will answer my research question.

I'm clueless at the moment.

Grouping of research questions
Mood:  accident prone
Now Playing: Young Love (Sonny James)
Topic: Research questions
So, had my supervisory session on Monday with James and Doug and it seems that the research questions that I developed weren't so good and I've got to get them prepared by next week.

They suggested that I group the research questions into headings but not sure how I'm going to do that and what that will solve but hopefully it might give me some fresh ideas!

Also James suggested that I look at the methodologies that I might use since these might feed into what my research question might be - because need to also decide my preliminary methodology by next week.

Further, I need to think about what kind of theoretical approach I will be taking since this will determine how I analyse my data. The only two theoretical approaches that I do know is the cognitive and situative and I was hoping to unified them as Anderson et al suggested, because I don't think I can be partial to one, because I think learning is social as well as mind thing.

Then I have to think about my definition of learning, Gill, I think, was saying her definition of learning was once someone knows something more than what they started out with (or something to that effect) - sure it is a good definition but not sure how well one can operationalize it :) i.e. measure that.