Regarding the poll on the last page -- I feel like understanding the methodology of their survey is the most important determinant of the poll's validity. Phrasing of the question, sample bias, etc are all factors that could invalidate it.
on the ACRI website they have both a press release regarding the poll as well as the actual polling research:
As someone with a background in stats and psychological/social surveying, it makes interesting reading to me.
Now, if we read the way the questions were laid out in the latter document, we can see that they sort of "lead on" from one to the next. It starts by asking if the individual is aware of the tensions in the ECS, and then goes on to steadily add the weight of "responsibility" to the individual by "pressuring" them to make the hypothetical decision as to whether Australia should commit.
I would argue this could be a case of bias, and it may have been worthwhile to conduct such questions in a more spaced out way. However I don't think any survey is perfect, and I do think the questions listed are written quite clearly. The last question especially, regarding the effect of conflict on Australian trade with China, is quite a specific one.
What neither document provides is the exact methodology and populace surveyed, but I'm going to assume that the methodology and sample was a valid one -- this is really STATS 101 and I doubt a think tank from a relatively respected university would dare to tamper with something so sacred and simple.
Now, assuming the poll was conducted in a valid fashion, then I think we should take the results for what they do say in their limited scope. There will always be people in the media who frame a result without sufficient context, and your everyday fellow will automatically make assumptions as well.
I think we need to take these results in context of the time period they were taken in: maybe in a year, the situation will have changed dramatically, or maybe the results would differ if the surveyor asked individuals to first do a week of research on the topic before hand.
Also, I want to add that the scenarios are deliberately limited in scope -- this is both a practical necessity due to the way surveys are carried out, but they may also provide a source of bias, as it could be argued (like by Samuraiblue) that a conflict limited to ECS could spill over into other areas, and that if those additional conflicts were included in the survey scenarios then the voting might have come out differently. I'm not sure if this is a fully legitimate criticism, however. On the one hand, the results could be easily (and incorrectly) flaunted as something akin to "Australia will remain neutral in any China-Japan/US conflict!!!" which would be a false generalization of the result. But on the other hand, that doesn't mean we should dismiss the result simply because it is limited in scope; I think it is still useful for the exact situation described in the survey. That is to say, the results can only be considered in a scope described by the scenario -- if it was only a conflict between China-Japan/US limited to the ECS, then Australians (note, not the Aus govt) may prefer to stay neutral. The survey doesn't tell us if Australians would stay neutral if the conflict was across an extended period, if it expands beyond ECS scope, nor the effect of media coverage, and how exactly the conflict unfolds, etc.
I also think an interesting question would have been to see how much Australians value trade with China, versus differing kinds of conflicts and tensions. That unfrotunately is difficult to survey, and what they do instead is simply say that Australians believe siding against China would cause their trade to drop.
IMHO the most telling statistic from this survey is that slightly over half of Australians were unaware of the tensions in the ECS, and even fewer are aware of the US commitment to those tensions. Therefore we need to consider all the other subsequent results based on the fact that these two key facts are not majority knowledge. That is to say, most of the sample surveyed probably are not very aware of the geopolitical situation (which would be expected, of course).
In other words, the results of this survey simply represent the opinions of that brief time period of the overall Australian population who probably lack a strong awareness of the geopolitical effect of the ECS disputes on their own country.
That leads to awkward questions about surveys, voting, democracy, and how much one should consider the opinions of a population who are not fully aware of a situation... but that's a massive can of worms better left untouched.
I think the biggest take away from this survey is that more surveys on the matter is needed, and preferably in more detail. I would be most interested to see how much Australians value trade with China as years pass, and just at what kind of conflict scenario involving their traditional allies that they would be willing to sacrifice it.