Within every academic domain, there exist certain conventions concerning writing. These may be formally defined, as in the style guides of individual journals, or informally adopted through common practice.
One such common practice is signposting your writing; saying what you're about to say, saying it, then saying what you just said. But common practice isn't always good practice, and while some subtle signposting can be useful, most of the time it's unnecessary, patronising to the reader and dull.
If you have a table of contents or an abstract, and if your writing flows smoothly from one point to the next, there's no need to outline every single step in advance and there's no need to say everything three times. Say it once and move on.
Steven Pinker, cognitive scientist and linguist, gives the following example in chapter 2 of "A Sense of Style"
The rest of this chapter is organised as follows. The first subsection introduces the concept of "metadiscourse", followed by one of its principal manifestations, the use of signposting. The second subsection reviews three issues: the problem of focusing on a description of professional activity rather than an exposition of subject matter, the overuse of apologetic language, and the disadvantages of excessive hedging. Following this, the third subsection explains the issue of prespecified verbal formulas. The fourth subsection covers issues having to do with excessive abstraction, including overuse of nominalization and passives. Finally, I will review the main points of the preceding discussion.
The problem here is that this paragraph only makes sense after you've read the rest of the chapter. It does not help the reader in any way, instead overloading them with information they can neither make sense of nor remember. As Pinker puts it...
The problem with thoughtless signposting is that the reader has to put more work into understanding the signposts than she saves in seeing what they point to... It's better if the route is clearly enough laid out that every turn is obvious when you get to it.
Well said. If the writing is good, signposting usually isn't necessary. If it's bad, signposting doesn't help.
Setting up expectations
As I'm writing this I'm faced with a choice. I could say, "the previous section stated that there are problems with signposting, the next will outline other approaches", but do I really need to tell you what I've just told you? No. And do I really need to tell you what's next? I don't think so. Far better to say something like, "if such explicit signposting is a problem, is there a better approach?" The signposting here is hidden within the rhetorical question, providing an implicit link between the previous point and the next.
Or I could strip it down further and just say, "is there a better way?". It's a natural question to ask, following on from the previous point, and asking the question implies that I'm going to try to answer it. As long as I meet the expectation I've just planted in the reader's mind, no further signposting is necessary.
Or I could strip away all traces of signposting and just say, "a better way is to...", in which case I wouldn't need to tell you where we're going because we've already arrived; answering the question before it's been asked.
Occasionally, cross referencing other chapters is useful. For example, if you say "this will be discussed further in chapter 6", it indicates that you have more to say about the subject. You can also refer back to previous sections, "as seen in chapter 3..."If you feel you need a paragraph setting out the scope of a chapter, by all means say something like, "the following sections will outline the primary mechanisms involved in nanostructure formation", but keep it as concise as possible. You don't have to list all the mechanisms in advance.
A couple of things to avoid...
When you start a new section, the heading is a sufficient signpost; you don't have to say what the previous section said; the reader knows because they just read it. If they don't, it's better to go back and edit the previous section.
I've read a few theses where every chapter begins with "the aim of this thesis is..."; if the reader gets to chapter 7 and has forgotten the point of the thesis, there's a problem!