Some conferences ask for an extended abstract. What are the differences among "abstracts," "extended abstracts," and "full papers?"
An abstract is a preliminary submission that summarizes the contribution of a paper. There are usually strict limits on the length of an abstract, either in terms of words or of total characters (rarely do they exceed 1000 words or 5000 characters; they are often substantially shorter than this.
An extended abstract and a full paper are nearly the same; the primary difference is that an extended abstract tends to be somewhat shorter than a full paper; I've seen extended abstracts from 2 pages up to 6 pages, while conference papers run from 4 up to about 12 or 15, depending on the space allotted.
One other important difference—outside of computer science, extended abstracts almost never go through a formal peer-review process before being published, while a conference paper will typically have at least one reviewer. (I think this is the case for almost all such papers, but there may be exceptions.)
In computer science conferences, the terms "full paper" and "extended abstract" are often used interchangeably.
For example, STOC calls it an "extended abstract" while SODA calls it a "full paper". In this case there is no difference that you can infer from the choice of the terminology: they are of the same length (approx. 10 pages + appendix) and they go through a similar peer review process.
Reading the other answers, it is definitely not the case that all disciplines understand the same when speaking about extended abstracts (while they can agree on abstracts and full papers).
In Economics, an extended abstract is something between an abstract and an introduction - being more close to the introduction. That is, it includes the What, the Why and a little bit of the the How, along with references and results. Results are, however, only preliminary, which is why there is only an extended abstract. There are usually no tables nor graphs in it. There are also no chapters.
I believe the Why is the most important part, as it distinguishes your work from the literature and shows what you are going to add.
My view is that in economics an extended abstract is a short version of the paper. Conferences that accept extended abstracts usually expect them to include results, methodology and a short discussion, to be able to gauge the plausibility of findings and appropriateness of the methodology.
In terms of lengths, I'd say 2-6 pages, while a full paper is anything from 15 to 100 pages. So, while very different in terms of length, the main contributions in the paper should be found in the extended abstract.