Zezima:
I'd try and avoid commenting on bias as it's not really something the author is consciously doing in order to persuade the audience, and that's really where the focus should be. However, bias can give way to some worthwhile devices, eg.
antanagogue which is similar to what you're describing. Analysing deliberate decisions the author has made is usually more effective than commenting on their actual opinion, as the latter can get a tad evaluative.
LiquidPaperz:
Again, I can't just give you the ideas, that's part of the task for Context. I've listed heaps on the examples thread as a starting point, but you're meant to be exploring these avenues on your own. What do you think a good idea might be? When you try and come up with a contention for the prompt, what are you basing your judgements on?
Refer to the link in first post regarding which forms/styles are 'best.'
Jesse C:
I'll give you some general comments now, but in future try and leave this thread for general questions. The Submissions and Marking Board is where you'd post full essays if you want feedback

-Try to paraphrase rather than quote absolutely anything that fits. Sentences like ' suggesting they ‘inhabit’ a ‘region’ of the ‘real world’,' sound a bit laboured, and it's really only the last two words that are important to your analysis.
-Re: overarching analysis, read the 'key players method' explanation earlier in this thread (should be a link in the first post.) This should give you a half-way point between really general contention discussion, and close/technique-based analysis.
-
It's not compulsory to mention a technique/device every time you analyse. In fact, a lot of what's persuasive in articles isn't the use of a rhetorical question or alliteration, but rather an overall appeal to one of our core values. Techniques are still worth mentioning of course, but don't let them be the basis for your analysis. Personally, I found it easier to work inwards from a big idea (ie. para focus: 'the author wants us to view the government as foolish and inept' --> what are some examples of this positioning --> how are the audience made to feel as a result --> why does the author want them to feel this way.) Each paragraph can do this what-how-why pattern a couple of times, but it should also have an overall sense of starting big, zooming in, and zooming back out to that big idea. This gives your analysis more direction, and even though structure isn't a huge part of the L.A. criteria, it can be easier for your assessor to follow a paragraph that begins with 'The author's aggressive demonisation of the government forms part of his attack on all who support the proposal' as opposed to 'The author uses inclusive language seven times, and also three rhetorical questions.'
Not all good analyses do it this way, but it should provide you with a solid option at least.
-Yes, it's okay to break up analysis sometimes; you don't want to be too formulaic in your approach. Usually some of the best and most insightful analysis comes when you examine how certain techniques work in conjunction with one another, or something like that. You should never evaluate the argument though, (ie. don't zoom out too far and end up talking about the issue, or pitting one approach against another in terms of which one is more persuasive.)
I'm sure I've missed at least one of your questions, but have a read through this and the key player explanation and get back to me, hopefully that should clear up some misconceptions.
edit: the reason you've heard conflicting recommendations is because every teacher has their own preferred method. Some like all the analysis spelt out for them, others prefer a very well structured piece, even if the analysis is shaky. Ideally you'll find your own variation that fulfills the criteria, so of course you can modify any approach as you see fit. Structuring by key players was something I discovered towards the end of Term 3, so not all of my analyses are like this. However, once I settled into that format, I found the actual process of analysing much more intuitive.
Yaccoub:
It's not an official criterion, and you could potentially score well without mentioning one or two, but if something has been included, then there's probably a reason, and I guarantee you'll fine at least one thing to analyse in each comment/text. I'd probably try and cover them all, unless you think doing so would negatively impact your other analysis. If you favour long-ish body paragraphs then you should be able to link one or two of them into the main text (either by common appeal (eg. '...this forms part of A's sympathetic appeal to parental concern. Similarly, B also calls upon readers' desire to protect the vulnerable, though his argument is much more defensive...'), or just the fact that their contentions are in opposition (eg. '...thereby furthering A's contention that football is awesome. By contrast, B refutes to sport as seen...') )
Possible structure for a core article with two images and three comments:
Para 1: Core text + first image
Para 2: Core text + first comment + second comment
Para 3: Core text + third comment + second visual
And you can mix and match depending on your paragraph numbers/length, as well as which comments work well when discussed together. Also, for most VCAA pieces, they'll tend to give you really obvious links between multi-text pieces. For instance, whatever the author is discussing during a powerpoint presentation will probably link to what's being displayed on the screen if there's a relevant image (this was especially true of 2010)
Valyria:
There's no real difference between 'Discuss' and 'Do you agree?' In both instances you'll be discussing whether or not you agree

Though I found, for certain texts at least, the 'Do you agree?'s invited more challenging than the 'Discuss' prompts. Rather than having just one paragraph dedicated to a challenge though, I tried to make my contention more complex and challenge throughout all the body paragraphs. This didn't always work though, so the 3:1 ratio is probably better than 4 paragraphs that ignore alternate views. Don't undercut yourself too much; the end of your challenge should come back to your contention, eg. '...thus Juliet also shouldered some responsibility. However, ultimately it was Romeo's foolhardy pursuit of love that unraveled both the lovers and their families.'
AmericanBeauty:
The idea of a tree falling and no one hearing it is pretty common, so you should be able to refer to the idea without the need to quote. Even if it's quite short, the evidence is usually more effective when you reword it to suit your purpose, rather than simply reciting a poem you've memorised.
yang dong:
Try and come up with some of your own ideas, even if you're worried they may be wrong, being able to critically think for yourself is essential at the time of year. Just start unpicking the idea irrespective of the text if you need; why is telling a story important, and why might those stories be inaccurate or embellished sometimes?
magneto:
thesaurus.com/browse/utmostthesaurus.com/browse/importance 
It seems like your evidence centres around the power of family as a unifying force, maybe start there? What exactly are you having trouble with in forming a contention? Giving you one to use isn't really teaching you anything.
Apologies if I've misread/left out parts of your questions. Let me know if you guys need any more help, but in general, please don't post anything along the lines of 'how do I answer this prompt' or 'what are some ideas for ___.' That might be what you need help with, but rather than using me or the forums as an idea generator, try and work out why you're having trouble. What is it about this theme or prompt that's thrown you off? At this stage, whilst using others can be invaluable in furthering your ideas, you have to have some foundation on which to build, and ultimately, you'll be on your own in the exam room, so developing a skillset for working through unfamiliar problems or concepts is more important than covering absolutely everything in a sophisticated way.
All the best!