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.

Met with James in the corridor
Mood:  hug me
Now Playing: I'm a Believer (The Monkees)
Topic: Meetings
So, just met James in the corridor and we were discussing the BSRLM conference and I told him although it was ok, I didn't think it was that useful. He suggested that sometime in future I should make a trip down to Bristol, since they are doing some interesting things there albeit all in schools!

Anyway, he said it will be good to dip into looking at conversation analysis (I told him I thought it looked hard :) ) - but he said I should still look at it, particularly look at Diane Laurillard book since she speaks about a framework for comparing software uses in difference disciplines (at least I think so) and if that gets more interesting to look at the work of the original developer Harvey Sacks.

So, will have to do that .. thought Gill had the book but couldn't find it on her shelf so may have to get it from the library.

Brainstorming about PhD research: Methodology
Mood:  chillin'
Now Playing: Travelin' Man (Rick Nelson)
Topic: Methodology
So, as part of the U500 session I attended on Tuesday we were also required to put together a list of tasks that we will undertake in the coming year. So, I listed a generic list of tasks, so here the are:

1. Literature Review
2. Decide on Methodology/ Methods
3. Negotiate Access with Universities
4. Pilot a study (still thinking about it!)
5. Observation of teaching LP in universities

Well, decided to expand on these just not to make it look so generic. So, first up is the literature review, which I'm looking at different types of software including example-based, problem-based and context-based as well as black-box and white-box software. Besides that my literature review may have to include something on cognitive learning (so far looked at a paper on cognitive load theory) and then something on computer-intensive curriculum and learning strategies with computers.

Deciding on the methodology or methods is difficult and I have no clue about that as yet. I was thinking I might have to do some kind of experimental design. The normal experiments that happen in schools is that the class is split and one group of students do one things and the other group do something different and then compare these groups - not sure how that is going to work - since I'll need access to a group of students to do that - and to keep the experiment of the two groups homogeneous somewhat such as same lecturer, same sort of approach (maybe depending on what I'm doing). Also, if I want to do it for several classes or disciplines, I'll also have to keep in mind that LP may be taught less or more in some disciplines or subjects as well for varying lengths sometimes in 2 hrs or 10 hrs!

I just scribbled something her on measuring student's usage of the software - not quite sure what I mean but perhaps how they use the software and what length etc., but then I was also considering to do a questionnaire of the student's attitude towards the differing software or I can do a pre-test/ post-test experimental type thing, give them both kinds of software and see what they think about it afterwards - mot quite sure if this will measure anything useful in this contest since giving a questionnaire before they even know LP and then giving them one after they know LP - how's that going to help anything - probably could get their attitude to what they expect the software to do for them and then see what it actually did - but this is all from their perception which may be 'coloured' and hence cannot provide actually what the software may/ may not have done for them - I guess that is why there is triangulation of methods in cases like these.

So, decided to group the pilot study and the observation study into one since I probably can do both in the same instance. So, for the pilot study I was thinking I can observe how students use the various software or how teachers employ the software and interview the students and teachers as to what they think about the using of the software etc. But before I can do this, probably, I should do the following:

1. Create a list of nearby universities with linear programming courses
2. Get the course lecturers contact details
3. Write to them to gain permission to come observe and interview them
4. Get a list of dates/ times when LP is being taught so you can travel and visit them

Thinking of PhD research: research questions
Mood:  a-ok
Now Playing: All I Have to Do is Dream (The Everly Brothers)
Topic: Research questions
So, went to the U500 session yesterday on research questions and design, so got me thinking about my research question, so I slapped down a few questions - they are definitely not perfect but needed to create one for the session: Well a very broad question I developed was:

"How do student learn LP using various types of software?"

This was quite broad but this is what I want to answer essentially, I do need to identify what is the types of software - so far, I have two definitions: one that goes with the box approach but this for solution software: so white-box and black-box, and then I can go to how they teach (software used in the teaching process) such as: problem-based software, example-based software or context-based (such as using problems that are related to their disciplines like agriculture, engineering, business). Well, using the box-approach to the classification of softwares, I developed a somewhat other question and this is based on Jonathan research (and somewhat what James is interested in):

What strategies do students employ when learning LP using a) black-box software and b) white-box software?

I'm not quite sure what I mean by strategy :) and what I hope to get from this. My third research question is alot like the first, so just will state it because I think I will throw it out soon:

How do students learn using software that are example-based, problem-based or context-based?

I was thinking about what James was saying about groups depending on their background may 'take to' differing software so perhaps something like this:

How does a student background influence their learning of LP using software?

But here I'm not sure what I mean by background - do I mean their disciplinary background? Their attitudes to computer? Their mathematical ability? Their age? Their sex? Their race? Their culture? All these are background factors now got to ask myself which do I mean! Perhaps, I can go back to what I originally wanted to study such as:

How does employing various types of LP software affect the learning of LP by students?

This is what I'm mainly interested in, and this suggests to me that the learning of LP may also depend on the course, the curriculum, the discipline etc and thus depends on how the software is integrated into the course and to what purpose it is used in the course. I expect in some cases where learning LP is done mostly by hand to learn everything, then employing a software to do the solutions only, wouldn't really have much of an effect. My problem comes here in what I consider learning of LP to be ... that is how do I measure it, and what constitutes that someone has learned LP. The question to ask is does LP consists of someone being taught the formulation of the problem, the solution of the problem (and which solutions these are: the simplex, the graphical, the interior-point?) and the sensitivity analysis (is this just learning the range of the coefficients of objective variables or the range of RHS variables or finding the dual price as well?). The next hard part is deciding when students have learnt LP is that when they are able to put the numbers into the software and get the answer out? Able to interpret the answer? Understand the theory behind it? Able to produce the answer by doing it by hand? Understand the principle? I think I have to soon (but not as yet since I don't want to limit my thinking)to make strong barriers and definitions for my research, for example, looking at courses teaching the graphical method only or the simplex algorithm only. Or looking at courses where the examination is by hand or something like that - got to think and decide about it. You may have note that my research question has nothing about disciplines in it ... and I do want to compare disciplines so somehow got to tuck that in - or let it be some parameter for comparison in my study.

Keith Trigwell and John's presentations
Mood:  a-ok
Now Playing: Sooner or Later (Supertramp)
Topic: Seminars
Thought the last entry was getting too long so decided to post it in a next entry.

Keith presentation although alright didn't engage my attention as much as Noel's (which is quite surprising since I was more interesting in Trigwell and Prosser work in the MSc!). I think perhaps because he was looking at students' perceptions and approached to learning from Oxford University and since I'm somewhat prejudice towards them decided to not pay much attention :). He talked mostly of the development of the model for student learning, particularly the 3P model (Presage, Process and Product) and how each part of the model relates to 'real' life such as degree result, deep and surface approach, teaching workload etc.

He did mention that in the Oxford undergraduate system, the humanities/social sciences tended to a lot more preparation for their work before a tutorial (and this perhaps increased their deep approach??). He also mentioned something called adoptive and adaptive learning by Blackmore which seems as something that might be of some interest to look up just to check what it is about.

Keith did have something interesting in his findings, that students had 1st class honours tended to perceive their workload as being appropriate and the same for good teaching. Which perhaps just means that students who are probably bright/ intelligent just think workload and teaching are good because the work is easy enough for them :)

They looked at the perception of students, their approaches to teaching and their motivation and conception. Keith was able to identify two clusters from these results. He found that for English, law, maths, history and physics students that the clusters were significantly different for each other for all the scales. However for engineering students the "good teaching" and "clear goals and standards" for perception was not significant between the clusters, and for the languages the deep approach clusters were not significantly different and for the Philosophy, Politics and Economics (PPE) students, the clear goals and standards were not significant either.

In John's presentation he employed the use of path analysis (wasn't quite sure how this was different to structural equation modelling - SEM - but found a chapter on it on the internet - it seems that SEM is an extension of path analysis which includes the latent variables). So, John decided to test the causal relationships between demographic background (age, sex etc), perceptions of academic environment, and study behaviour for predicting outcome measures (he used marks I believe for this one) by controlling the effects of each of the predictor variables, and test whether they were acting as mediating variables. His path analysis obviously assumes that these are the only variables that affect outcome measures. In the end his results suggest that perceptions of academic environment and study behaviour both individually act to predict (or cause) the outcome measures and also, that both the perceptions of academic environment and study behaviour act through one another to cause the outcome measure.

This finding was not wholly accepted by Keith as one should expect since his framework suggested that Perceptions of academic environment and study behaviour was acting together to make the reality or perspective of a student and does not have a causal relationship between them.

In the discussion forum, John, Keith or Noel mentioned that Meyer was putting together an ASI bibliography for group disciplines. This sounded as something interesting to get my hands on.

I think this session was really good, just got to meet Ramsden and Meyer and would have met all of the famous people from my MSc research :).